Chapter 1. State of the Art 1.1 Interaction of Millimeter Waves with Biological Matter

THESE DE DOCTORAT DE

L'UNIVERSITE DE RENNES 1 COMUE UNIVERSITE BRETAGNE LOIRE

ECOLE DOCTORALE N° 601 Mathématiques et Sciences et Technologies de l'Information et de la Communication Spécialité : Electronique

présentée par Rosa ORLACCHIO

Millimeter waves for biomedical electromagnetics: study of changes induced at the cellular level by pulsed electromagnetically-induced heating

Thèse présentée et soutenue à Rennes, le 4 Juillet 2019 Unité de recherche : IETR – UMR CNRS 6164, Institut d’Electronique et des Télécommunications de Rennes – Université de Rennes 1

Rapporteurs avant soutenance :

Theodoros SAMARAS Professor, Aristotle University of Thessaloniki Liliana Delia ARNAUD-CORMOS Maître de Conférences, Université de Limoges – XLIM

Composition du Jury :

Examinateurs : Theodoros SAMARAS Professor, Aristotle University of Thessaloniki Liliana Delia ARNAUD-CORMOS Maître de Conférences, Université de Limoges – XLIM Micaela LIBERTI Assistant Professor, University of Rome "La Sapienza" Ronan SAULEAU Professeur, Université de Rennes 1, directeur de thèse

Invité(s) Maxim ZHADOBOV Chargé de Recherche CNRS, IETR, co-directeur de thèse Yves LE DREAN Maître de Conférences, Université de Rennes 1

This work was supported by the French Agency for Food, Environmental and Occupational Health and Safety (ANSES) through NEAR 5G and MEMO projects; the EU through the European Regional Development Fund (ERDF), Brittany Region, Ministry of Higher Education and Research, Rennes Métropole and Conseil Départemental, through the CPER Project SOPHIE / STIC & Ondes.

Contents

Resumé...... i Abstract...... xi General introduction...... 1 Chapter 1. State of the art 1.1 Interaction of millimeter waves with biological matter ...... 11

1.1.1 Biological targets of millimeter waves ...... 11 1.1.2 Electromagnetic and thermal properties of the skin ...... 15 1.1.3 General features of interaction of millimeter waves with the human body ...... 22 1.1.4 Tissue-equivalent models ...... 29

1.2 In vitro studies at millimeter waves ...... 33

1.2.1 Exposure systems ...... 33 1.2.2 Biological effects of millimeter waves ...... 34 1.2.3 Numerical dosimetry ...... 37 1.2.4 Experimental dosimetry ...... 38 1.2.5 Techniques for measurement of MMW-induced heating ...... 38

1.3 Heat-induced biological effects: from continuous to pulsed-modulated waveforms 44

1.3.1 Thermal treatments ...... 44 1.3.2 Thermal dose ...... 52 1.3.3 Importance of pulsed waveform in heat-induced cellular responses ...... 60

Bibliography...... 67 Chapter 2. Electromagnetic and thermal dosimetry at millimeter waves: effects of thermal convection during in vitro exposure of cells

2.1 Introduction...... 83 2.2 Heat transfer and convection ...... 84

2.2.1 Conduction ...... 84 2.2.2 Radiation ...... 85 2.2.3 Convection...... 87

2.3 Experimental set-up ...... 91

2.4 Samples under test ...... 94

2.4.1 Electromagnetic properties ...... 96 2.4.2 Thermo-physical properties ...... 96

2.5 Temperature measurements ...... 97

2.5.1 Operation principle of thermocouple ...... 98 2.5.2 Impact of thermocouple on temperature measurements...... 99 2.5.3 Measurements of thermal pulses ...... 105 2.5.4 Thermocouple configuration ...... 107 2.5.5 Reproducibility ...... 108

2.6 Electromagnetic dosimetry ...... 112

2.6.1 Impact of cell monolayer ...... 120

2.7 Temperature–based retrieval of SAR ...... 122 2.8 Continous wave ...... 126

2.8.1 Heating in different media ...... 126 2.8.2 Dependence on SAR ...... 130 2.8.3 Role of liquid volume ...... 133

2.9 Pulse-induced heating ...... 135

2.9.1 Dependence on pulse duration ...... 135 2.9.2 Heating in water and agar gel ...... 138

2.10 Conclusions ...... 140

Bibliography...... 142 Chapter 3. Cell mortality and heat shock stress induced by pulsed modulated millimeter wave heating at cellular level

3.1 Introduction...... 147 3.2 Biological models ...... 149

3.2.1 Cell line ...... 150 3.2.2 Exposure protocol...... 151 3.2.3 Immunocytochemistry protocol ...... 153 3.2.4 Image acquisition and fluorescence analysis ...... 155 3.2.5 Statistical analysis ...... 157

3.3 Electromagnetic and thermal dosimetry ...... 158

3.3.1 1.5 s heat pulses ...... 159 3.3.2 6 s heat pulses ...... 162 3.3.3 Thermal dose ...... 167

3.4 Heat shock response ...... 168

3.4.1 Dose-response curves for continuous wave heating ...... 170 3.4.2 Pulsed vs continuous heating ...... 173

3.5 Cellular apoptosis ...... 178

3.5.1 Theoretical basis of cellular apoptosis ...... 179 3.5.2 Pulsed vs continuous heating ...... 182

3.6 Discussion ...... 189 3.7 Conclusions ...... 192 Bibliography ...... 193 General conclusions ...... 200 Future work and perspectives ...... 202 Appendices

A Millimeter-wave interaction with eyes ...... 207 B Measurement of complex permittivity ...... 211 C Measurement of thermal properties ...... 217 D Measurement of viscosity ...... 225

About the author ...... 229 List of publications ...... 230 Acknowledgement ...... 231

Résumé

La discipline à l'interface entre la biologie et l'électromagnétisme est appelée bioélectromagnétisme. Elle vise à évaluer i) les mécanismes d'interaction des ondes électromagnétiques avec le corps humain, ii) les bénéfices potentiels découlant de l'exploitation de différentes parties du spectre électromagnétique à des fins thérapeutiques, et iii) les risques associés à l'exposition des êtres humains aux champs électromagnétiques. De nombreuses approches ont été définies pour atteindre ces objectifs, notamment des études théoriques et analytiques, des expériences in vivo et in vitro, ainsi que des études en laboratoire ou épidémiologiques sur des humains. L'exploration de la partie haute du spectre microondes, plus particulièrement la bande des ondes millimétriques (OMM), a suscité un intérêt croissant au cours de 20 dernières années en raison du développement rapide des nouvelles technologies dans cette bande [1]. Les fréquences des OMM se situent entre 30 et 300 GHz, correspondant à une longueur d'onde λ en espace libre de 10 à 1 mm, respectivement. Les OMM sont des rayonnements non ionisants car l'énergie photonique (0.12–1.24 meV) reste inférieure, de plusieurs ordres de grandeur, au niveau requis pour ioniser les molécules biologiques (typiquement > 10 eV). De plus, l'énergie associée aux OMM n'est pas suffisante pour casser des liaisons chimiques, d'hydrogène ou des liaisons van der Waals, toutes impliquées dans l'interaction entre molécules biologiques au sein des cellules. Seule l'énergie de rotation des dipôles moléculaires (10-5–10-2 eV) est inférieure à l'énergie photonique des OMM, ce qui signifie que ce type de rayonnement peut interagir avec des molécules qui ont un degré de liberté de rotation, comme les molécules d'eau. Par conséquent, suite à l'interaction avec des OMM, l'énergie thermique des molécules d'eau augmente, ce qui entraîne un échauffement de l'eau ou des tissus à forte teneur en eau (c.-à-d. des tissus mous comme la peau). Les OMM sont absentes du spectre électromagnétique naturel en raison de la forte absorption résonnante du rayonnement induit par l'oxygène dans l'atmosphère (16 dB/km) autour de 60 GHz [2], [3]. Par conséquent, le vivant n’a jamais interagi avec ces rayonnements dans des conditions naturelles. L'utilisation croissante des technologies en OMM a suscité un intérêt considérable en termes des effets potentiels de ces rayonnements sur le vivant, stimulant de nombreuses recherches visant à expliquer les mécanismes d'interaction avec le corps humain et à comprendre les risques potentiels découlant de l'utilisation de cette bande [4]–[7]. La saturation du spectre microondes et la disponibilité de larges bandes à des fréquences plus élevées ont stimulé le développement des systèmes de communication en bande millimétrique. Par rapport aux dispositifs fonctionnant à des fréquences plus basses, la bande millimétrique offre de

i nombreux avantages, notamment des débits plus élevés (au-delà de plusieurs Gbit/s), la taille plus compacte des antennes, et des interférences réduites avec des dispositifs voisins [8]. Nombreuses applications sans fil bénéficient des OMM, y compris les réseaux locaux sans fil (WLAN) [9], les réseaux personnels sans fil (WPAN) [10], les réseaux corporels (BAN) [11], et les réseaux émergents 5G qui devraient être déployés à partir de 2022 [12], [13]. D'autres applications des OMM incluent des radars d'assistance à la conduite automobile [14], des scanners de sécurité aéroportuaire [15], et des armes non létales de contrôle de foule [16]. Diverses applications dans le domaine biomédical peuvent également bénéficier des OMM, notamment les dispositifs de surveillance non invasive du glucose dans le sang [17], [18], l'évaluation de l'hydratation cornéenne [19], la détection des caries dentaires [20], et l'imagerie médicale pour le diagnostic des cancers de la peau [21], [22] et du sein [23], [24]. En outre, les OMM de faible puissance (densité de puissance incidente 10 ≤ mW/cm2), à des fréquences spécifiques comprises entre 40 et 60 GHz (généralement 42.2, 53.6, et 61.2 GHz) ont été appliquées seules ou en combinaison avec d'autres traitements (dans certains pays d'Europe de l'Est), comme modalité thérapeutique dans des dispositifs médicaux pour guérir des diffèrent maladies [25], [26], notamment le mélanome sous cutané [27], [28]. Il a été démontré que les OMM de faible puissance peuvent avoir des effets médicaux sur les réponses analgésique [29] et inflammatoire [30], [31], ce qui suggère que le rayonnement millimétrique peut interférer avec des processus cellulaires. Cependant, malgré les résultats cliniques prometteurs, aucune preuve claire de mécanismes impliqués n'est disponible à ce jour. La faible profondeur de pénétration des OMM suggère que les effets biologiques des OMM pourraient être amorcés dans la peau par les terminaisons nerveuses du système nerveux périphérique. L'absorption de l'énergie en OMM produit une élévation de la température dans les tissus exposés, agissant en fonction de la dose. Si l'intensité d'exposition est ≤ 5 mW/cm2, l'élévation de température produite dans les tissus est ≤ 0.5°C [1], alors que les expositions de forte puissance (supérieure à plusieurs centaines de mW/cm2) peuvent induire une élévation significative de la température entraînant une sensation de douleur ou des dommages tissulaires [1]. L'existence d'effets purement électromagnétiques, à faible intensité d'exposition, qui ne dépendent pas directement de l'élévation de température, reste encore controversée [1]. L'application des OMM, dans le domaine biomédical décrit ci-dessus, est actuellement limitée aux dispositifs de faible puissance ne générant qu'un échauffement imperceptible des tissus exposés. Cependant, l'intensité des OMM de forte puissance, induisant un échauffement important de la peau, pourraient être exploitée pour le traitement thermique des cancers superficiels de la peau.

L’augmentation de la température induite par les ondes électromagnétiques a été largement utilisée comme moyen de traitement de cancers [32]. Les fréquences micro-ondes (MW) ont été exploitées pour des traitements thermiques invasifs et non invasifs, principalement dans les bandes appelées Industrial Scientific Medical (ISM) autour de 434 MHz, 915 MHz, et 2.45 GHz. De nombreux avantages des fréquences plus élevées, tels qu'une résolution spatiale plus élevée et la taille compacte des structures rayonnantes, ont motivé l'exploration de la bande MW jusqu'à 18 GHz pour l'ablation locale des tumeurs avec une invasivité minimale et des dommages collatéraux réduits (9.2 GHz [33], 10 GHz [34], 14.5 GHz [35], et 18 GHz [36]). Récemment, il a été démontré analytiquement [37] que la partie basse de la bande millimétrique (c.-à-d. 20–100 GHz) peut être utilisée pour une focalisation spatialement précise de la chaleur à l'intérieur de la peau, en variant la fréquence et la taille du faisceau, ainsi qu'en appliquant une convection d'air pour réduire la surchauffe de la peau, en déplaçant la zone d’élévation maximale de la température vers des couches plus profondes. Ces résultats suggèrent que la bande 20–100 GHz représente une alternative intéressante pour le traitement thermique non invasif du cancer de la peau, comme le mélanome, situé à l'intérieur de la profondeur de pénétration des OMM (c.-à-d. environ 0.5 mm dans la peau à 60 GHz). De plus, il a été démontré que, contrairement à l’échauffement continu qui caractérise l'hyperthermie conventionnelle, des impulsions thermiques courtes peuvent permettre d'atteindre rapidement une température crête élevée dans la région d'intérêt, ce qui peut entraîner des dommages plus importants au niveau des cellules ciblées [38], tout en évitant de léser les tissus sains entourant la tumeur [39]. Les ultrasons (par ex. les ultrasons focalisés à haute intensité pulsés (p-HIFU) [40]) et différentes bandes du spectre électromagnétique, comme les radiofréquences (RF) [41], MW [38], et les rayonnements optiques (laser) [39], ont été étudiés jusqu'ici comme une source impulsionnelle d’échauffement pour l'ablation du cancer. Toutefois, l'utilisation de fréquences plus élevées, comme des OMM, reste encore inexplorée à cette fin. L'équipe WAVES de l'Institut d'Electronique et de Télécommunications de Rennes (IETR) a récemment proposé une technique précise pour chauffer localement des monocouches de cellules en utilisant des impulsions d’OMM modulées en amplitude (PW) [42]. À l'aide d'un système d'exposition ad hoc en OMM, des impulsions thermiques d'amplitude et de durée différentes peuvent être générées localement in vitro. Une méthodologie détaillée a été proposée [42] pour mesurer à l'échelle microscopique l’échauffement induit par les OMM, permettant de retrouver le débit d'absorption spécifique (DAS) à 58.4 GHz. Cependant, les événements thermodynamiques, tels que la convection thermique et les gradients de température élevés dans le milieu de culture cellulaire, n’ont été jamais étudiés en détail. Ces événements thermodynamiques sont absents dans

iii les échantillons non exposés utilisés comme contrôle, ce qui peut entraîner une mauvaise interprétation des résultats biologiques observés. Ce projet de recherche doctoral a été motivé par ces avancées récentes [37], [42], ainsi que par l'intérêt croissant pour le traitement thermique des tumeurs utilisant des impulsions thermiques générées par des OMM. L'objectif principal de cette étude est d’évaluation des modifications potentielles induites dans les cellules de mélanome malin exposées à 58.4 GHz en régime continu et impulsionnel, avec la même augmentation moyenne de température. La réponse au choc thermique médié par la phosphorylation d'une protéine de choc thermique de faible poids moléculaire (HSP27) et l'activation d'une voie apoptotique par la Caspase-3, a été évaluée au moyen d'analyse d'images par microscopie à fluorescence. Cette étude interdisciplinaire à l'interface électromagnétisme / biologie cellulaire a été réalisée à l'IETR en étroite collaboration avec l'équipe TrEC (Transcription, Environnement et Cancer) de l'Institut de Recherche sur la Santé, l’Environnement et le Travail (IRSET, Inserm UMR_S 1085, Université de Rennes 1), représentée par le Dr Yves Le Dréan et le Dr Yann Le Page, et avec le Dr Stanislav Alekseev de l'Institut de Biophysique Cellulaire (ICB, Puschino, Russie).

Plan de thèse et nouvelles contributions Le manuscrit est divisé en trois chapitres. Le premier chapitre dresse un bilan précis de l'état de l'art. Les principales nouveautés et contributions de cette thèse sont décrites en détail aux chapitres 2 et 3, qui portent respectivement sur les aspects physiques (dosimétrie électromagnétique et thermique) et biologiques (analyse des réponses cellulaires spécifiques au chauffage induit par les OMM). Le manuscrit est organisé comme suit. Le Chapitre 1 définit l'état de l'art dans ce domaine. Ce chapitre est divisé en trois sections. La première section décrit l'interaction des OMM avec la peau, principale cible biologique des rayonnements millimétriques. En particulier, la structure de la peau est détaillée et ses propriétés électromagnétiques et thermiques sont présentées. Des modèles de peau homogènes et multicouches sont illustrés pour représenter les interactions électromagnétiques et thermiques avec les OMM. Des fantômes équivalents des tissus, simulant les propriétés électromagnétiques de la peau et d'autres tissus en bande MMW, sont décrits. La deuxième section décrit l'état actuel des connaissances sur les études in vitro dans la bande millimétrique, y compris les systèmes d'exposition, les effets biologiques, la dosimétrie numérique et expérimentale, et les techniques de mesure de l’échauffement induit par les OMM. La troisième section décrit les connaissances actuelles sur les thérapies thermo-oncologiques, telles que l'hyperthermie et l'ablation thermique, les réponses

iv cellulaires et moléculaires à la chaleur, les modèles de la dose thermique, et les avantages possibles découlant de l'utilisation des impulsions thermiques pour le traitement du cancer. Le Chapitre 2 porte sur la dosimétrie électromagnétique et thermique du système d'exposition spécifique utilisé lors des expériences in vitro. En particulier, les travaux présentés dans ce chapitre visent à étendre les connaissances actuelles sur la dosimétrie thermique in vitro dans la bande millimétrique, au problème spécifique de la convection qui est souvent négligé dans les expériences bioélectromagnétiques. Pour la première fois, l'impact de la convection a été évalué expérimentalement en détail en OMM. En particulier, nous avons étudié les effets de la convection sur la dynamique de la température dans un scénario typique d'exposition in vitro lors d'un échauffement induit à 58.4 GHz en continu et en régime impulsionnel. À l'aide de plusieurs modèles, dont l'eau distillée, un milieu de culture, et un gel d’agarose (modèle équivalent à l’eau mais sans convection), les effets de la convection sur l’échauffement local ont été étudiés en comparant la cinétique de température enregistrée localement dans les liquides à celle du gel sans convection. Différents paramètres, qui peuvent varier d'une expérience in vitro à l'autre, ont été analysés, notamment i) la viscosité, ii) le DAS, iii) le volume de liquide, et iv) la durée de l'impulsion thermique. Nos résultats suggèrent que dans la planification des études in vitro, la convection, déclenchée par la distribution particulièrement non-uniforme du DAS en OMM, est un paramètre très important. Elle peut avoir un impact indirect sur la réponse cellulaire par le biais de nombreux effets qui sont absents dans le contrôle, y compris le changement local de la concentration en oxygène ou en nutriments transportés par le milieu de culture. Cette étude a révélé plusieurs caractéristiques importantes du développement de la convection dans les liquides exposés à 58.4 GHz: 1) l'apparition de la convection modifie la dynamique de température du milieu exposé, 2) une monocouche de cellules d’une épaisseur comprise entre 5–10 μm n'a pas d'impact sur la dynamique de la température dans le milieu de culture et n'influence donc pas le déclenchement de la convection, 3) le début de la convection est précédé d'un pic de température, dont l’apparition dans le temps dépend du DAS (plus le DAS est élevé, plus la convection se déclenchée tôt), 4) l'apparition de la convection dépend du volume du liquide (des volumes plus importants provoquent un déclenchement plus précoce de la convection et un effet de refroidissement plus important), 5) le déclenchement de la convection dépend de la viscosité du fluide (des valeurs plus élevées de la viscosité du fluide provoquent une convection plus faible), et 6) la convection modifie fortement la forme de l'impulsion de chaleur en diminuant la température de pointe et en augmentant la vitesse de refroidissement. Le Chapitre 3 traite de la réponse aux chocs thermiques des cellules de mélanome malin suite à une exposition à 58.4 GHz en continu et en régime impulsionnel, avec la même augmentation de

v température moyenne. Pour la première fois, la réponse hyperthermique des cellules de mélanome a été analysée à différentes durées d'impulsion, à savoir 1.5 s et 6 s, tant en termes d'induction de phosphorylation de HSP27, que d'activation de la Caspase 3. Le chapitre commence par une description détaillée des matériaux et des méthodes utilisés, y compris la lignée cellulaire, le protocole d'exposition, l'acquisition des images, et l'analyse des données. Cette section est suivie par la présentation de la dosimétrie électromagnétique et thermique. Successivement, la réponse au choc thermique quantifiée en suivant la phosphorylation de HSP27, comme indicateur du stress cellulaire, est présentée et analysée. La base théorique des mécanismes biologiques impliquant la phosphorylation de HSP27 est décrite. Enfin, l'activation de la Caspase-3 par clivage a été évaluée comme un indicateur de l'apoptose cellulaire, avec une tentative de corréler le stress thermique aux dommages thermiques létaux dans les cellules. Les voies intrinsèques et extrinsèques de l'apoptose sont également décrites de manière non exhaustive au début de la section. Les résultats ont démontré que l’exposition continue avec une température inférieure à 43°C, correspondant au seuil thermique typique capable d'induire la mort dans de nombreuses lignées cellulaires pour une exposition prolongée [43], n'est pas suffisant pour induire l'apoptose cellulaire dans les cellules du mélanome malin, et ne convient donc pas au traitement thermo-oncologique des cancers cutanés superficiels. La distribution de la phosphorylation de HSP27 suggère que la thermotolérance a été induite dans les cellules exposées en continue, entraînant l'inhibition efficace de l'activation de la machine apoptotique. Au contraire, des impulsions thermiques de haute intensité avec une amplitude crête de 10°C ont induit une réponse cellulaire plus forte que celle induite par l’échauffement continu, tant en termes de phosphorylation de HSP27 que d'activation de la Caspase-3. En particulier, dans la zone jusqu'à 1.8 mm du centre du puits, l'apoptose cellulaire observée après l'exposition aux impulsions thermiques a montré une forte augmentation, compatible avec l’hypothèse que le système de "surveillance qualité des protéines" est saturé. En effet, la présence dans la même région du puits, d'un plateau dans la phosphorylation des HSP27 suggère que l’activation des HSP27 a déjà atteint son maximum, empêchant toute prise en charge supplémentaire des protéines soumises à un stress thermique, entraînant de ce fait la mort cellulaire. De courtes impulsions thermiques de 1.5 s ont produit une plus grande réponse que des impulsions de 6 s. Plus précisément, dans la zone située entre 0 et 0.5 mm de l'axe du guide d'onde utilisé comme source, la phosphorylation de HSP27 induite par des impulsions thermiques de 1.5 s était 1.4 et 1.2 fois supérieure à celle induite par des impulsions de 6 s et par l’exposition continue, respectivement. L'activation correspondante de la Caspase-3 induite par des impulsions de chaleur de 1.5 s était 2.3 et 5.2 fois supérieure à celle induite par des impulsions de 6 s et par l’exposition continue, respectivement. La durée de l'impulsion thermique détermine l'extension de la région du

vi stress thermique induit par la phosphorylation de la HSP27. Cela signifie que plus la durée de l'impulsion thermique est élevée, plus la chaleur est diffusée à l'intérieur du puits exposé, déterminant une plus grande extension de la zone de contrainte. Les résultats obtenus dans cette étude ouvrent la voie à une nouvelle stratégie de destruction des cellules de mélanome à faible échauffement moyen. Le manuscrit se termine par une Conclusion qui résume les principaux résultats et contributions, ainsi que les travaux futurs et perspectives. La thèse contient quatre Annexes. L'Annexe A décrit l'état de l'art de l'interaction des OMM avec les yeux. L'Annexe B décrit la procédure expérimentale utilisée pour mesurer la permittivité complexe de l'eau, du milieu de culture, et du gel d’agarose à l'aide d'une sonde coaxiale ouverte.

L'annexe C décrit la procédure utilisée pour mesurer la capacité calorifique spécifique cp (J/(kg∙°C)) et la conductivité thermique k (W/(m∙°C)) de l'eau et du milieu de culture avec un calorimètre différentiel à balayage et un analyseur de conductivité thermique, respectivement. Enfin, l'annexe D décrit la procédure utilisée pour mesurer les valeurs de la viscosité de l'eau et du milieu de culture.

vii

Bibliographie [1] M. Zhadobov, N. Chahat, R. Sauleau, C. Le Quement, and Y. Le Drean, “Millimeter-wave interactions with the human body: state of knowledge and recent advances,” Int. J. Microw. Wirel. Technol., vol. 3, no. 02, pp. 237–247, Apr. 2011. [2] F. Giannetti, M. Luise, and R. Reggiannini, “Mobile and personal communications in the 60 GHz band: a survey,” Wirel. Pers. Commun., vol. 10, no. 2, pp. 207–243, Jul. 1999. [3] A. W. Straiton, “The absorption and reradiation of radio waves by oxygen and water vapor in the atmosphere,” IEEE Trans. Antennas Propag., vol. 23, pp. 595–597, Jul. 1975. [4] A. G. Pakhomov, Y. Akyel, O. N. Pakhomova, B. E. Stuck, and M. R. Murphy, “Current state and implications of research on biological effects of millimeter waves: a review of the literature,” Bioelectromagnetics, vol. 19, no. 7, pp. 393–413, 1998. [5] A. Ramundo-Orlando, “Effects of millimeter waves radiation on cell membrane - A brief review,” J. Infrared Millim. Terahertz Waves, vol. 31, no. 12, pp. 1400–1411, Dec. 2010. [6] Y. Le Dréan et al., “State of knowledge on biological effects at 40–60 GHz,” Comptes Rendus Phys., vol. 14, no. 5, pp. 402–411, May 2013. [7] S. Alekseev and M. Ziskin, “Biological effects of millimeter waves and submillimeter waves,” in Biological and medical aspects of electromagnetic fields, Fourth Edn., Boca Raton, FL: CRC Press, 2018. [8] M. Zhadobov, C. Leduc, A. Guraliuc, N. Chahat, and R. Sauleau, Antenna/human body interactions in the 60 GHz band: State of knowledge and recent advances. Institution of Engineering and Technology, 2016. [9] K. Sakaguchi et al., “Millimeter-wave wireless LAN and its extension toward 5G heterogeneous networks,” ArXiv150704518 Cs, Jul. 2015. [10] E. Callaway et al., “Home networking with IEEE 802.15.4: a developing standard for low- rate wireless personal area networks,” IEEE Commun. Mag., vol. 40, no. 8, pp. 70–77, Aug. 2002. [11] Q. H. Abbasi, M. U. Rehman, K. Qaraqe, and A. Alomainy, Advances in body-centric wireless communication: applications and state-of-the-art. IET Digital Library, 2016. [12] T. S. Rappaport et al., “Millimeter wave mobile communications for 5G cellular: It will work!,” IEEE Access, vol. 1, pp. 335–349, 2013. [13] A. R. Guraliuc, M. Zhadobov, R. Sauleau, L. Marnat, and L. Dussopt, “Near-field user exposure in forthcoming 5G scenarios in the 60 GHz band,” IEEE Trans. Antennas Propag., vol. 65, no. 12, pp. 6606–6615, Dec. 2017. [14] J. Hasch, E. Topak, R. Schnabel, T. Zwick, R. Weigel, and C. Waldschmidt, “Millimeter- wave technology for automotive radar sensors in the 77 GHz frequency band,” IEEE Trans. Microw. Theory Tech., vol. 60, no. 3, pp. 845–860, Mar. 2012. [15] D. M. Sheen, D. L. McMakin, and T. E. Hall, “Three-dimensional millimeter-wave imaging for concealed weapon detection,” IEEE Trans. Microw. Theory Tech., vol. 49, no. 9, pp. 1581– 1592, Sep. 2001. [16] C. Yu and M. Cui, “Numerical study of deposition of energy of active denial weapon in human skin,” 2012 Asia-Pac. Symp. Electromagn. Compat., pp. 361–364, 2012. [17] P. H. Siegel, Y. Lee, and V. Pikov, “Millimeter-wave non-invasive monitoring of glucose in anesthetized rats,” in 2014 39th International Conference on Infrared, Millimeter, and Terahertz waves (IRMMW-THz), 2014, pp. 1–2. [18] Y. Nikawa and T. Michiyama, “Blood-sugar monitoring by reflection of millimeter wave,” in 2007 Asia-Pacific Microwave Conference, 2007, pp. 1–4. [19] D. B. Bennett et al., “Assessment of corneal hydration sensing in the terahertz band: in vivo results at 100 GHz.,” J. Biomed. Opt., vol. 17, no. 9, pp. 97008–1, 2012.

viii

[20] N. Hoshi, Y. Nikawa, K. Kawai, and S. Ebisu, “Application of microwaves and millimeter waves for the characterization of teeth for dental diagnosis and treatment,” IEEE Trans. Microw. Theory Tech., vol. 46, no. 6, pp. 834–838, Jun. 1998. [21] F. Topfer and J. Oberhammer, “Millimeter-wave tissue diagnosis: the most promising fields for medical applications,” IEEE Microw. Mag., vol. 16, no. 4, pp. 97–113, May 2015. [22] A. Mirbeik-Sabzevari, S. Li, E. Garay, H.-T. Nguyen, H. Wang, and N. Tavassolian, “Synthetic ultra-high-resolution millimeter-wave imaging for skin cancer detection,” IEEE Trans. Biomed. Eng., vol. 66, no. 1, pp. 61–71, May 2018. [23] S. D. Meo et al., “On the feasibility of breast cancer imaging systems at millimeter-waves frequencies,” IEEE Trans. Microw. Theory Tech., vol. 65, no. 5, pp. 1795–1806, May 2017. [24] I. Iliopoulos et al., “Enhancing breast cancer imaging at millimeter waves using focusing techniques,” in 2017 11th European Conference on Antennas and Propagation (EUCAP), 2017, pp. 1017–1019. [25] M. Rojavin, “Medical application of millimetre waves,” QJM, vol. 91, no. 1, pp. 57–66, Jan. 1998. [26] M. C. Ziskin, “Physiological mechanisms underlying millimeter wave therapy,” in Bioelectromagnetics Current Concepts, 2006, pp. 241–251. [27] A. A. Radzievsky, O. V. Gordiienko, I. Szabo, S. I. Alekseev, and M. C. Ziskin, “Millimeter wave-induced suppression of B16 F10 melanoma growth in mice: involvement of endogenous opioids,” Bioelectromagnetics, vol. 25, no. 6, pp. 466–473, Sep. 2004. [28] I. Szabo et al., “Destruction of cutaneous melanoma with millimeter wave hyperthermia in mice,” IEEE Trans. Plasma Sci., vol. 32, no. 4, pp. 1653–1660, Aug. 2004. [29] T. I. Usichenko, H. Edinger, V. V. Gizhko, C. Lehmann, M. Wendt, and F. Feyerherd, “Low-intensity electromagnetic millimeter waves for pain therapy,” Evid. Based Complement. Alternat. Med., vol. 3, no. 2, pp. 201–207, Jun. 2006. [30] V. R. Makar, M. K. Logani, A. Bhanushali, S. I. Alekseev, and M. C. Ziskin, “Effect of cyclophosphamide and 61.22 GHz millimeter waves on T-cell, B-cell, and macrophage functions,” Bioelectromagnetics, vol. 27, no. 6, pp. 458–466, Sep. 2006. [31] V. R. Makar, M. K. Logani, A. Bhanushali, M. Kataoka, and M. C. Ziskin, “Effect of millimeter waves on natural killer cell activation,” Bioelectromagnetics, vol. 26, no. 1, pp. 10– 19, Jan. 2005. [32] I. Mellal, A. Oukaira, E. Kengene, and A. Lakhssassi, “Thermal therapy modalities for cancer treatment: a review and future perspectives,” Int. J. Appl. Sci. - Res. Rev., vol. 4, no. 4, Nov. 2017. [33] D. A. Hodgson, I. B. Feldberg, N. Sharp, N. Cronin, M. Evans, and L. Hirschowitz, “Microwave endometrial ablation: development, clinical trials and outcomes at three years,” BJOG Int. J. Obstet. Gynaecol., vol. 106, no. 7, pp. 684–694, Jul. 1999. [34] H. Luyen, F. Gao, S. C. Hagness, and N. Behdad, “Microwave ablation at 10.0 GHz achieves comparable ablation zones to 1.9 GHz in ex vivo bovine liver,” IEEE Trans. Biomed. Eng., vol. 61, no. 6, pp. 1702–1710, Jun. 2014. [35] C. P. Hancock, N. Dharmasiri, M. White, and A. M. Goodman, “The design and development of an integrated multi-functional microwave antenna structure for biological applications,” IEEE Trans. Microw. Theory Tech., vol. 61, no. 5, pp. 2230–2241, May 2013. [36] J. Yoon et al., “High-frequency microwave ablation method for enhanced cancer treatment with minimized collateral damage,” Int. J. Cancer, vol. 129, no. 8, pp. 1970–1978, Oct. 2011. [37] M. Zhadobov, S. I. Alekseev, Y. Le Dréan, R. Sauleau, and E. E. Fesenko, “Millimeter waves as a source of selective heating of skin,” Bioelectromagnetics, vol. 36, no. 6, pp. 464– 475, Sep. 2015. [38] M. Bedoya, A. M. del Rio, J. Chiang, and C. L. Brace, “Microwave ablation energy delivery: Influence of power pulsing on ablation results in an ex vivo and in vivo liver model,” Med. Phys., vol. 41, no. 12, Dec. 2014. ix

[39] M. Ganguly, S. Miller, and K. Mitra, “Model development and experimental validation for analyzing initial transients of irradiation of tissues during thermal therapy using short pulse lasers,” Lasers Surg. Med., vol. 47, no. 9, pp. 711–722, Nov. 2015. [40] B. E. O’Neill, H. Vo, M. Angstadt, K. P. C. Li, T. Quinn, and V. Frenkel, “Pulsed High Intensity Focused Ultrasound mediated nanoparticle delivery: mechanisms and efficacy in murine muscle,” Ultrasound Med. Biol., vol. 35, no. 3, pp. 416–424, Mar. 2009. [41] B. Zhang, M. A. J. Moser, E. M. Zhang, Y. Luo, and W. Zhang, “Numerical analysis of the relationship between the area of target tissue necrosis and the size of target tissue in liver tumours with pulsed radiofrequency ablation,” Int. J. Hyperthermia, vol. 31, no. 7, pp. 715– 725, Oct. 2015. [42] M. Zhadobov, S. I. Alekseev, R. Sauleau, Y. Le Page, Y. Le Dréan, and E. E. Fesenko, “Microscale temperature and SAR measurements in cell monolayer models exposed to millimeter waves,” Bioelectromagnetics, vol. 38, no. 1, pp. 11–21, Jan. 2017. [43] A. S. Sapareto and W. C. Dewey, “Thermal dose determination in cancer therapy,” Int. J. Radiat. Oncol., vol. 10, no. 6, pp. 787–800, Apr. 1984.

Abstract

The lower part of the millimeter wave (MMW) band (i.e., 20–100 GHz) is an attractive alternative for non-invasive thermal treatment of skin cancer such as melanoma. Besides, pulsed electromagnetically-induced heating can lead to stronger damage in cells compared to traditional continuous heating. In in-vitro experiments, continuous-wave (CW) or pulsed-wave (PW) amplitude-modulated MMW can be efficiently used to locally heat cell monolayers with a typical thickness ranging between 3 µm and 10 µm. In this work we investigate the potential modifications induced at the cellular level in melanoma cells following exposure to CW and PW MMW-induced heating with the same average temperature rise, at 58.4 GHz. First, the impact of thermal convection on temperature dynamics in models representing typical in vitro exposure conditions during CW and PW-induced heating is experimentally investigated. To this end, we analyse electromagnetic and thermophysical parameters impacting convection and varying from one in vitro experiment to another, namely SAR, viscosity, liquid volume, and duration of a thermal pulse. Second, the heat shock response, mediated by phosphorylation of a small heat shock protein (HSP27) and activation of Caspase-3, indicator of cellular apoptosis, are quantified to monitor the biological response using an experimental approach based on fluorescence microscopy. Two pulse durations (1.5 s and 6 s) are considered. Our results demonstrate that thermal pulses are able to induce a stronger cellular response in melanoma cells both in terms of heat shock and cellular mortality compared to the one induced by CW. The shorter the pulse duration, the greater the cellular response.

Keywords: Millimeter waves, convection, in vitro experiments, pulsed-induced heating, heat shock response, non-invasive destruction of cells.

General introduction

The interaction of electromagnetic field with biological matter has been the subject of interest for the last 70 years. The need for understanding the effects of the electromagnetic field on the human body emerged in the 1940s and 1950s, possibly stimulated by the rapid increase in research on radiofrequency technologies after the World War II, when scientists began questioning the safety levels of exposure of humans working with high-power radars or devices exploiting radiofrequencies in medicine and industry [1]. The interdisciplinary field at the interface between biology and electromagnetics is referred as bioelectromagnetics. It is aimed to assess i) the mechanisms of interaction of the electromagnetic waves with the human body, ii) the potential benefits deriving in exploiting different parts of the electromagnetic spectrum for therapeutic purposes, and iii) the risk associated to exposure of humans to the electromagnetic fields. Numerous approaches have been defined to achieve these goals, including theoretical and analytical studies, in vivo and in vitro experiments, and human laboratory or epidemiological investigations. The exploration of the upper part of the microwave spectrum, namely the MMW band, has received an increasing interest in the last 20 years due to the rapid development of new technologies in this band [2]. The MMW region of the electromagnetic spectrum refers to the extremely high frequency (EHF) band between 30–300 GHz, corresponding to a free-space wavelength λ from 10 to 1 mm, respectively. MMW are non-ionising radiations as the photon energy (0.12–1.24 meV) remains several orders of magnitude below the level required to ionise biological molecules (typically > 10 eV). Besides, the energy associated with MMW is not sufficient to break ionic, hydrogen or van der Waals bonds, all involved in the interaction between biological molecules within cells. Only the rotational energy molecular dipoles (10-5–10-2 eV) is lower than the photon energy of MMW, meaning that this type of radiation may interact with molecules that have a rotational degree of freedom, such as water molecules. Therefore, following the interaction with MMW, the thermal energy of the water molecules raises, resulting in the heating of water or tissues with a high water content (i.e., soft tissues such as skin). MMW radiations are absent in the environmental electromagnetic spectrum because of the strong resonant oxygen-induced radiation absorption in the atmosphere (16 dB/km) around 60 GHz [3], [4], meaning that biological species have never interacted with such radiations in natural conditions. The increasing use of MMW technologies created considerable interest in potential effects of those radiations on biological systems, stimulating large amount of research aiming to explain the mechanisms of interaction with the human body and to understand the possible risks

General introduction deriving from the use of such band for the establishment of accurate safety standard guidelines [5]– [8]. The saturation of the spectrum currently exploited for civil telecommunications and the availability of larger channel bandwidths at higher frequencies boosted the development of communication systems at MMW. Compared to devices operating at lower frequencies, the MMW band offers many advantages, including faster data rates (beyond several Gbit/s), more compact size of radiating structures, and lower interferences between devices [9]. It is upcoming for numerous wireless applications, including wireless local area networks (WLAN) [10], wireless personal area networks (WPAN) [11], body area networks (BAN) [12], and emerging 5G networks, expected to be available by 2020 [13], [14]. Other applications of the MMW band include automotive driver assistance radars[15], airport security scanners [16], and non-lethal crowd control weapons [17], [18]. Various applications in the biomedical field also may benefit from the MMW spectrum including devices for non-invasive blood glucose monitoring [17], [18], corneal hydration assessment [19], dental caries detection [20], and medical imaging for skin [21], [22] and breast cancers diagnosis [23], [24]. In addition, low-power MMW (incident power density ≤ 10 mW/cm2), at specific frequencies between 40 and 60 GHz (typically 42.2, 53.6, and 61.2 GHz) have been applied alone or in combination with other treatments, in some Eastern European countries as a therapeutic modality in medical devices to cure diseases [25], [26], including subcutaneous melanoma [27], [28]. It was demonstrated that low-power MMW may have medical effects on both analgesic [29] and inflammatory [30], [31] responses, suggesting that MMW radiation may interfere with several cellular processes. However, despite of the promising clinical results, no clear evidence of involved mechanisms is available thus far. The shallow penetration of the MMW radiation suggests that MMW bioeffects must be initiated within the skin through the nerve endings of the peripheral nervous system. Heating is induced by the absorption of the electromagnetic energy in the 10–100 GHz range [2]. Absorption of MMW energy produces a temperature elevation in the exposed tissue, acting in a dose-dependent manner. If the exposure intensity is ≤ 5 mW/cm2, the temperature rise produced in the exposed tissue is ≤ 0.5°C [2], while high-power exposures (above several hundreds of mW/cm2) may induce significant temperature elevation leading to a pain sensation or tissue damage [2]. The existence of purely electromagnetic effects, at low-intensity exposures, that do not directly depend on temperature rise remains still controversial [2]. The application of MMW, in the biomedical field, is currently limited to low-power devices generating only imperceptible heating of the exposed tissues. However, high-power MMW

General introduction intensity, inducing substantial heating of the skin, may be exploited for the thermal treatment of superficial skin cancers. Electromagnetically induced heating has been largely used as a means to treat numerous types of cancers [32]. Microwave (MW) frequencies have been exploited for invasive and non-invasive thermal treatments, mainly in the Industrial Scientific Medical (ISM) bands around 434 MHz, 915 MHz, and 2.45 GHz. Numerous advantages of higher frequencies, such as higher spatial resolution and compact size of radiating structures, motivated exploration of MW up to 18 GHz for local tumor ablation with minimized invasiveness and collateral damages (9.2 GHz [33], 10 GHz [34], 14.5 GHz [35], and 18 GHz [36]). Recently, it has been analytically demonstrated [37] that the lower part of the MMW band (i.e., 20–100 GHz) can be employed for spatially-accurate focusing of heat inside the skin by varying frequency and exposure beam size, as well as by enforcing air convection to reduce overheating of skin surface shifting the maximum heating towards deeper skin layers. Results suggested that the 20–100 GHz range represents an attractive alternative for non- invasive thermal treatment of skin cancer such as spreading melanoma, located within the MMW penetration depth (i.e., around 0.5 mm in the skin at 60 GHz). Moreover, it was demonstrated that in contrast to continuous heating, which characterizes conventional hyperthermia, short thermal pulses may provide the rapid achievement of high peak temperature in the area of interest, possibly leading to stronger damage of cells [38], eluding the injury of the healthy tissue surrounding the tumor [39]. Ultrasound (i.e., pulsed-High Intensity Focused Ultrasound (p-HIFU) [40]) and different bands of the electromagnetic spectrum, such as radiofrequency (RF) [41], MW [38], and light (laser) [39] have been investigated so far as a source of pulsed thermal heating for cancer ablation. However, the use of higher frequencies, such as the MMW band, remains still unexplored for this purpose. WAVES team of the Institut d’Electronique et de Télécommunication de Rennes (IETR) has recently proposed an accurate technique to locally heat cell monolayers by means of high-intensity pulsed wave (PW) amplitude modulated MMW [42]. Using an ad hoc MMW exposure system, heat pulses of different amplitude and duration may be locally generated in vitro. A detailed methodology has been also proposed [42] for the microscale measurement of MMW-induced heating and consequent specific absorption rate (SAR) retrieval at the level of a cell monolayer in culture medium exposed at 58.4 GHz. However, thermodynamic events, such as thermal convection, generated by the high SAR and temperature gradients within the cell culture medium has nether been investigated in detail so far. These thermodynamic events are absent in the non- exposed samples used as control, possibly leading to misinterpretation of the observed biological results.

General introduction

This PhD research project is motivated by these recent advances [37], [42] as well as by the increasing interest in pulsed thermal treatment of tumors. The main goal of this PhD study is to investigate potential modifications induced in malignant melanoma cells exposed at 58.4 GHz to continuous (CW) and PW MMW-induced heating with the same average temperature rise. The heat shock response mediated by the phosphorylation of a low molecular weight heat shock protein (HSP27) and the activation of an apoptotic pathway through Caspase-3, has been evaluated by means of fluorescence microscopy image analyses. The dissertation will firstly position the work in the current state of the art. Then electromagnetic and thermal dosimetry is presented, with the main focus on the impact of thermal convection on temperature dynamics in models representing typical in vitro exposure conditions during CW and PW-induced heating. Finally, modifications induced by heat pulses of different durations in melanoma cells in vitro are presented and discussed. This interdisciplinary study at electromagnetics / cell biology interface has been carried out at IETR in close collaborations with the TrEC (Transcription, Environnement et Cancer) team of the Institut de Recherche sur la Santé, l’Environnement et le Travail (IRSET, Inserm UMR_S 1085, Université de Rennes 1) represented by Dr. Yves Le Dréan and Dr. Yann Le Page, and Dr. Stanislav Alekseev from the Institute of Cell Biophysics (ICB, Puschino, Russia).

Novel contributions and thesis outline The manuscript is divided into three chapters. The first chapter overviews the state of the art. The main novelties and contributions of this work are detailed in Chapter 2 and Chapter 3, focused on the physical (i.e., electromagnetic and thermal dosimetry) and biological (i.e., analysis of the specific cellular responses to MMW-induced heating) aspects, respectively. More in detail, the manuscript is organized as follows. Chapter 1 overviews the state of the art in the field. The chapter is divided into three sections. The first section describes the interaction of the MMW with skin, the main biological target of the MMW radiation. In particular, skin structure is detailed, and its electromagnetic and thermal properties are presented. Both homogeneous and multi-layered models of skin are illustrated to represent the electromagnetic and thermal interactions with the MMW. Tissue-equivalent phantoms simulating the electromagnetic properties of skin and other tissues in the MMW range are described. The second section overviews the current state of the art on in vitro studies at MMW including exposure systems, induced biological effects, numerical and experimental dosimetry, and techniques for the measurement of the MMW-induced heating. The third section describes the current knowledge about thermo-oncological therapies, such as hyperthermia and thermal ablation, 4

General introduction cellular and molecular responses to heat, mathematical modeling of thermal dose, and possible advantages deriving from the use of pulsed waveform in heat induced damage. Chapter 2 focuses on the electromagnetic and thermal dosimetry for the specific exposure system employed during the in vitro experiments. In particular, the work presented in this chapter aims to extend the current knowledge about in vitro thermal dosimetry at MMW to the specific problem of thermal convection, which is usually neglected in MMW studies. For the first time, the impact of the thermal convection has been experimentally evaluated in detail at MMW. In particular, we studied the effects of thermal convection on temperature dynamics in a typical in vitro exposure scenario during CW and PW MMW induced heating at 58.4 GHz. Using several models, including distilled water, culture medium, and agar gel (convection-free water-equivalent model), the effects of convection on local heating have been investigated by comparing temperature kinetics locally recorded in the liquids to that in the water-equivalent gel without convection. Different parameters, which may differ from one in vitro experiment to another, have been analyzed including i) viscosity, ii) SAR, iii) liquid volume, and iv) duration of a thermal pulse. Chapter 3 deals with the heat shock response of malignant melanoma cells following exposure to CW and PW amplitude modulated MMW-induced heating with the same average temperature rise at 58.4 GHz. For the first time, the hyperthermic response of melanoma cells was analysed at different pulse durations, namely 1.5 s and 6 s, both in terms of induction of phosphorylation of HSP27 and activation of Caspase 3. Results obtained in this study pave the way towards a new strategy for destruction of melanoma cells at low average heating. The chapter begins with a detailed description of the materials and methods employed in the study, including the cell line, exposure protocol, picture acquisition, and data analysis. This section is followed by the presentation of electromagnetic and thermal dosimetry. Successively, the heat shock response quantified by following the phosphorylation of HSP27, as an indicator of cellular stress, is presented and analyzed. Finally, Caspase-3 cleaved activation was evaluated as an indicator of cellular apoptosis, with the attempt to correlate the thermal stress to the effective heat damage in cells. Theoretical basis of the biological mechanisms involving the phosphorylation of HSP27 as well as the activation of an apoptotic pathways via Caspase-3 are non-exhaustively illustrated at the beginning of the relative section. The manuscript is concluded with Conclusion that summarizes the most significant results and major contributions as well as identifies perspectives for future development both from the physical and biological points of view. The dissertation contains four Appendices. Appendix A reviews the state of the art in interaction of MMW with eyes. Appendix B describes the experimental procedure used to measure

General introduction the complex permittivity of water, culture medium, and agar gel using an open-ended coaxial probe DAK-1.2E (SPEAG, Zurich, CH). Appendix C describes the experimental procedure used to measure the specific heat capacity and thermal conductivity k (W/(m∙°C)) of water and culture medium with a differential scanning calorimeter (DSC, Q200, TA Instruments) and a thermal conductivity analyzer (C-THERM, Tci), respectively. Finally, Appendix D describes the experimental procedure used to measure the viscosity values of water and culture medium.

Bibliography

[1] W. Beck, “Bioelectromagnetics society, BEMS history book,” http://www.bioelectromagnetics.org/doc/bems-history.pdf. [Online]. Available: http://www.bioelectromagnetics.org/doc/bems-history.pdf. [2] M. Zhadobov, N. Chahat, R. Sauleau, C. Le Quement, and Y. Le Drean, “Millimeter-wave interactions with the human body: state of knowledge and recent advances,” Int. J. Microw. Wirel. Technol., vol. 3, no. 02, pp. 237–247, Apr. 2011. [3] F. Giannetti, M. Luise, and R. Reggiannini, “Mobile and personal communications in the 60 GHz band: a survey,” Wirel. Pers. Commun., vol. 10, no. 2, pp. 207–243, Jul. 1999. [4] A. W. Straiton, “The absorption and reradiation of radio waves by oxygen and water vapor in the atmosphere,” IEEE Trans. Antennas Propag., vol. 23, pp. 595–597, Jul. 1975. [5] A. G. Pakhomov, Y. Akyel, O. N. Pakhomova, B. E. Stuck, and M. R. Murphy, “Current state and implications of research on biological effects of millimeter waves: a review of the literature,” Bioelectromagnetics, vol. 19, no. 7, pp. 393–413, 1998. [6] A. Ramundo-Orlando, “Effects of millimeter waves radiation on cell membrane - A Brief Review,” J. Infrared Millim. Terahertz Waves, vol. 31, no. 12, pp. 1400–1411, Dec. 2010. [7] Y. Le Dréan et al., “State of knowledge on biological effects at 40–60 GHz,” Comptes Rendus Phys., vol. 14, no. 5, pp. 402–411, May 2013. [8] S. Alekseev and M. Ziskin, “Biological effects of millimeter waves and submillimeter waves,” in Biological and medical aspects of electromagnetic fields, Fourth Edn., Boca Raton, FL: CRC Press, 2018. [9] M. Zhadobov, C. Leduc, A. Guraliuc, N. Chahat, and R. Sauleau, Antenna/human body interactions in the 60 GHz band: State of knowledge and recent advances. Institution of Engineering and Technology, 2016. [10] K. Sakaguchi et al., “Millimeter-wave wireless LAN and its extension toward 5G heterogeneous networks,” ArXiv150704518 Cs, Jul. 2015. [11] E. Callaway et al., “Home networking with IEEE 802.15.4: a developing standard for low- rate wireless personal area networks,” IEEE Commun. Mag., vol. 40, no. 8, pp. 70–77, Aug. 2002. [12] Q. H. Abbasi, M. U. Rehman, K. Qaraqe, and A. Alomainy, Advances in body-centric wireless communication: applications and state-of-the-art. IET Digital Library, 2016. [13] T. S. Rappaport et al., “Millimeter wave mobile communications for 5G cellular: It will work!,” IEEE Access, vol. 1, pp. 335–349, 2013. [14] A. R. Guraliuc, M. Zhadobov, R. Sauleau, L. Marnat, and L. Dussopt, “Near-field user exposure in forthcoming 5G scenarios in the 60 GHz band,” IEEE Trans. Antennas Propag., vol. 65, no. 12, pp. 6606–6615, Dec. 2017. [15] J. Hasch, E. Topak, R. Schnabel, T. Zwick, R. Weigel, and C. Waldschmidt, “Millimeter- wave technology for automotive radar sensors in the 77 GHz frequency band,” IEEE Trans. Microw. Theory Tech., vol. 60, no. 3, pp. 845–860, Mar. 2012. [16] D. M. Sheen, D. L. McMakin, and T. E. Hall, “Three-dimensional millimeter-wave imaging for concealed weapon detection,” IEEE Trans. Microw. Theory Tech., vol. 49, no. 9, pp. 1581– 1592, Sep. 2001. [17] P. H. Siegel, Y. Lee, and V. Pikov, “Millimeter-wave non-invasive monitoring of glucose in anesthetized rats,” in 2014 39th International Conference on Infrared, Millimeter, and Terahertz waves (IRMMW-THz), 2014, pp. 1–2. [18] Y. Nikawa and T. Michiyama, “Blood-sugar monitoring by reflection of millimeter wave,” in 2007 Asia-Pacific Microwave Conference, 2007, pp. 1–4. [19] D. B. Bennett et al., “Assessment of corneal hydration sensing in the terahertz band: in vivo results at 100 GHz.,” J. Biomed. Opt., vol. 17, no. 9, pp. 97008–1, 2012.

Chapter 1 State of the art

Content

1.1 Interaction of millimeter waves with biological matter ...... 11

1.2 In vitro studies at millimeter waves ...... 33

1.3 Heat-induced biological effects: from continuous to pulsed-modulated waveforms 44

1.3.1 Thermal treatments ...... 44 1.3.2 Thermal dose ...... 52 1.3.3 Importance of pulsed waveform in heat-induced cellular responses ...... 60

Bibliography...... 67

CHAPTER 1. State of the art

1.1 Interaction of millimeter waves with biological matter

The aim of this section is to give an overview of the current state of the art related to different aspects of the interaction of the MMW with the skin, mainly from the electromagnetic and thermal points of view. The section is organized as follows. First, the biological targets of the MMW radiations are introduced and skin structure is described in detail. Second, electromagnetic and thermal properties of skin are presented as a foundation for the characterization of the interaction with MMW. Third, different models of skin (i.e., homogeneous or multi-layered) typically used to represent the electromagnetic and thermal interaction with the MMW are illustrated. Finally, tissue- equivalent models realized to simulate the electromagnetic properties of skin and other tissues in the MMW range are described.

1.1.1 Biological targets of millimeter waves Shallow penetration depth of MMW into biological tissues and solutions (of the order of a fraction of a millimeter, e.g., 0.34 mm at 58.4 GHz) indicates that skin or near-surface zones of the body, such as the eyes, are the main targets of the MMW radiation. Here, a detailed description of the skin structure is firstly provided. Interaction of MMW with eyes is non-exhaustively reviewed in Appendix A.

Skin organization The skin is the largest organ of the human body covering almost the entire surface and 15% of the body weight [1]. It has several functions including 1) protection against UV light, infections or other external harmful influences, 2) sensation of the external stimuli through many receptors, 3) thermoregulation through sweat glands, hair, and adipose tissues, and 4) metabolic functions such as production of vitamin D. Skin structure has been extensively studied with many complete references available in the literature such as [2]. It is made up by three main layers of different depth, namely epidermis, dermis, and subcutaneous fat layer (Fig. 1.1). The epidermis is the outer layer of the skin made up by 4 or 5 layers (depending on the body location): 1) stratum corneum, 2) stratum lucidum, 3) stratum granulosum, 4) stratum spinosum, and 5) stratum basale. The stratum lucidum is present only in the palms of the hands or the soles of the feet in order to provide a greater protection to those parts. The epidermis provides the protection of the body against potentially hazardous environmental threats through physical, chemical, 11

CHAPTER 1. State of the art

biochemical, and adaptive immunologic barriers [3]; it also provides mechanical resistance. The stratum corneum is the outermost layer of the epidermis whose typical thickness is of about 10– 25μm [4]. It is mainly composed of non-viable hexagonal flat cells called coenocytes and it plays a key role in maintaining the structural integrity and hydration of the skin. The rest of the epidermis (i.e., epidermis minus stratum corneum) is called viable epidermis whose thickness varies roughly between 30–100μm [5]. Four main groups of cells compose the viable epidermis: i) keratinocytes, which constitute about the 80% of the cells of the epidermis, migrate from the basal layer outwards to the skin surface, ii) melanocytes, responsible for the production of the pigment melanin and its transfer to keratinocyte, iii) Langerhans cells, involved in a variety of immune responses, and iv) Merkel cells, located in the stratum basale, working as mechanoreceptors essential for light touch sensation. The epidermis contains no blood-vessels and for this reason it completely depends on dermis for nutrient delivery and waste disposal. The dermis lies beneath the epidermis and is 20 to 30 times thicker than the epidermis, typically between 1.5 and 4 mm thick [6]. This layer contains most of the specialized cells and structures of the skin, including blood and lymph vessels, hair follicles, sweat glands, sebaceous glands, nerve endings, fibroblasts, collagen, and elastin. The main functions of the dermis are the regulation of the body temperature and the supply of the epidermis with nutrient-saturated blood. It is divided into two anatomical regions: 1) the papillary layer, containing a thin arrangement of collagen fibers and a thin extensive vascular system used both to provide nutrients to the epidermis and to regulate the temperature, and 2) the reticular layer made of thick collagen fibers, to strengthen the skin, providing structure and elasticity. This layer also supports other components of the skin, such as hair follicles, sweat glands, and sebaceous glands. The hypodermis (also called subcutaneous fat layer) is a layer directly below the dermis; it serves to connect the skin to the underlying fascia (fibrous tissue) of the bones and muscles. It protects the body from external trauma, insulates it from cold, and acts as energy reservoir thanks to its adipocytes. The blood vessels, nerves, lymph vessels, and hair follicles cross also this layer. The thickness of the hypodermis varies largely throughout the body and from person to person depending from age, sex, race, endocrine, and nutritional status of the individual. Typical thickness for non-obese adults ranges from 1.1 to 7.1 mm, while in obese or very obese adults it may vary between 10 to 30 mm [7].

CHAPTER 1. State of the art

Figure 1.1: Detailed structure of skin: (1) hair shaft; (2) stratum corneum; (3) sweat-pore; (4) hair follicle; (5) arrector pili muscle; (6) sebaceous gland; (7) nerve; (8) eccrine sweat gland; (9) cutaneous vascular plexes; (10) adipose depot. Subplots on the top represent the structure and cellular composition of (A) epidermis and (B) dermis [8].

Skin cancer Skin cancer represents one of the most common form of malignancy in the white population [9], [10]. The most common types of skin cancers are non-melanoma skin cancers (NMSCs) such as basal cell carcinoma (BCC) and squamous cell carcinoma (SCC). However, the risk associated to those cancers is typically low and the 90% of NMSC cases are successfully cured. On the other side, melanoma skin cancer, although is the least common (about 1% of all cases), represents the most deadly skin cancer. The incidence rate of melanoma and NMSC is increasing worldwide [11]. For example, the American Cancer Society has estimated that in 2019 there will be 96480 new cases of melanoma in the United States and 7230 deaths from the disease [12]. There are different types of melanoma including i) superficial spreading melanoma, ii) nodular melanoma, iii) lentigo maligna melanoma, iv) acral lentiginous melanoma, and v) amelanotic melanoma. According to the Melanoma Research Foundation (MRF, [13]), the superficial spreading

CHAPTER 1. State of the art

melanoma is the most common type accounting for approximately 70% of all diagnosed against the 5 to 15% due to the others. The most hazardous problem related to melanoma is its facility to spread from the superficial layers of skin toward the deepest ones. The risk profile and the extent of melanoma spread are described as staging [14], [15]. The most widely staging system used for the melanoma classification is the Tumor-Node-Metastasis (TNM) system provided by the American Joint Committee on Cancer (AJCC). Cancer staging systems are revised regularly to reflect the current medical knowledge available. The most recent revision of the cutaneous melanoma staging system is the 8th edition of the Cancer Staging Manual updated to the 1st January 2018 [16]. T refers to the classification of the cancer considering its thickness and ulceration, N to the involvement of lymph nodes (collections of cells that act as a filter for the immune system), M to whether there is any evidence for the cancer to spread to other parts of the body such as other skin sites, the lung, or other organs (M1), or not (M0). Once the T, N, and M categories are assigned, it is possible to determine the stage of the melanoma [14] (Fig. 1.2), e.g., stage 0 refers to superficial melanoma while stage IV refers to its invasion to other parts of the body.

Figure 1.2: Skin cancer stages [17].

CHAPTER 1. State of the art

1.1.2 Electromagnetic and thermal properties of the skin The characterization of the electromagnetic properties of the skin is fundamental to accurately assess the field behavior close to or inside the human body. In turn, the definition of the thermal properties of the skin is important to predict the temperature increase within the body following the MMW exposure.

Electromagnetic properties

The response of biological material to MMW exposure depends on its electric and magnetic properties. In biological non-ferromagnetic materials, magnetic response is very weak, and -7 permeability μ is very close to that of free space (μ0), i.e., μ ≈ μ0 = 4π·10 (H/m). The electric properties of material are characterized by the frequency-dependent complex permittivity:

(1.1) where is the real part of the complex permittivity (or dielectric constant) representing the induced polarization, and is the imaginary part (or dielectric loss) which is a measure of the friction associated with changing polarization and the drift of conduction charges [18]. The dielectric loss is related to the total conductivity σ by:

(1.2)

where σ is the total conductivity (S/m), ω=2πf (rad/s) is the angular frequency of the field (with f, the frequency (Hz)), and is permittivity of the free space ( ). According to Foster and Schwan [19], the loss factor can be also expressed as:

(1.3)

where is the magnitude of the dispersion of the free water fraction of the skin, is the optical permittivity, is the permittivity at <<1, is the an angular frequency, τ is the relaxation time (s), is the ionic conductivity of the skin, and is permittivity of the free space. The total conductivity σ includes a contribution from the frequency independent ionic

CHAPTER 1. State of the art

conductivity and is associated with the frequency-dependent dielectric loss of material. The contribution of ionic conductivity decreases with increasing frequency, and at MMW the total conductivity is defined mainly by the dielectric losses of material. The loss tangent is a measure of how “lossy” or energy absorbing a material is; it is defined as the ratio of the imaginary to real part of the complex permittivity:

(1.4)

The frequency dependence of the complex permittivity of the skin may be described by the Debye equation with a single (or multiple) relaxation time τ (s) equal to that of the pure water:

(1.5)

where is the optical permittivity, is the magnitude of the dispersion of the free water fraction of the skin, is the permittivity at << 1, is an angular frequency, τ is the relaxation time (s), and is the ionic conductivity of the skin. However, the complexity of both the structure and composition of the skin is such that each dispersion region may be broadened by multiple contributions to it. The broadening of the dispersion could be empirically accounted for by introducing a distribution parameter. The resulting equation is known as Cole-Cole equation:

(1.6)

where 0 <α < 1 is a measure of the broadening of the dispersion. The permittivity of the biological tissues is a function of the frequency and it generally decreases as frequency increases. The dependence of the permittivity of the material on the frequency of an applied electric field is defined dielectric dispersion [20]. The dielectric spectrum of biological tissues is characterised by three main major dispersion regions accounting for different mechanisms: α is attributed to the ionic diffusions, β to the capacitive charging of cellular membranes and intercellular bodies, and γ to the dipolar polarization of the free water molecules of the tissue. A fourth additional and weak dispersion named δ may be also distinguished between the β and γ ones and it appears in some proteins solutions [21], [22].

CHAPTER 1. State of the art

Figure 1.3: Electromagnetic properties (real part of the relative permittivity and conductivity σ) and relaxation regions of typical biological tissue as function of frequency.

The relaxation frequency of the γ dispersion in skin is about 22 GHz at 37°C [23], meaning that the γ dispersion is exhibited in the MMW frequency range. The main absorber of the MMW in skin is water [18]. The photon energy associated to the propagation of MMW is lower than both ionization and thermal motion energies; however it is higher than the energy of orientation motion of the molecular dipoles (10−3–10−4 eV). Thus, MMW can interact with molecules that have rotational degree of freedom such as water molecules. The result is the increase of the thermal energy of water molecules with the consequent generation of heating. Some of the water molecules are motionally restricted (i.e., "bound" water) due to the interaction with soluble tissue component forming a hydration shell. The remaining water molecules are free in motion and compose the “free” water component of tissue. The water of hydration associated with proteins and other organic molecules exhibits little absorption in the MMW frequency range due to strong restriction of their motion. Therefore the total amount of water contributing to MMW absorption in tissues equals the total water content minus the fraction of "bound" or motionally restricted water [24]. Electromagnetic properties of the skin have been well characterized in the frequency range below 20 GHz; however available data above 50 GHz are very limited [25]–[31] due to technical difficulties associated with measurements [32]. In addition, the results reported in the literature thus far strongly depend on the measurement technique, the sample type (in vivo or in vitro study), and other experimental conditions such as skin temperature, location on the body, and thickness of different skin layers. In the 10–100 GHz range, the dispersive dielectric properties of the skin and biological solutions are primarily related to the rotational dispersion of free water molecules. In particular, high losses are related to the free water relaxation with the peak at 23 GHz at 33°C [28]. Skin water content varies as a function of age, location on the body, gender, ethnic origin, etc [7]. In [25], an extrapolation from the measured data on skin rabbit at 23 GHz using a Debye model is performed at 60 GHz. In vitro studies on human skin samples have been performed in [26], and

CHAPTER 1. State of the art

the dielectric properties were obtained using a free space quasi-optical technique in the 60–100 GHz frequency range. In [27], Gabriel et al., reported the complex permittivity of the human skin up to 110 GHz, based on the extrapolation of measured data performed for frequencies up to 20 GHz on in vivo human skin. Two skin models have been identified, namely wet skin and dry skin. Hwang et al., measured in vivo the permittivity of human epidermis including the palm and the wrist using broadband measurement up to 110 GHz by means of a one mm-diameter coaxial probe [31]. Alekseev and Ziskin [28] performed in vivo reflection measurements using an open-ended waveguide and proposed homogeneous and multilayer human skin models by fitting the experimental data. Chahat et al., used an open-ended coaxial probe [29] and a new temperature- based method to measure human skin electromagnetic properties [30]. In the latter method [30], forearm skin was exposed to 60.4 GHz using an open-ended waveguide and continuous wave signal. Temperature distribution was recorded using an infrared camera. By fitting the analytical solution of the bio-heat transfer equation to the experimental heating kinetics, the values of the PD and penetration depth (δ) were found and used to retrieve the complex permittivity of skin described by the Debye equation (Eq. (1.5)). Table 1.1 provides a summary of the complex permittivity values at 60 GHz. Data are divided into in vitro or in vivo where the first refers to freshly excised tissues, while the latter to measurements carried out on accessible parts of the body. E and M stand for extrapolation and measurement, respectively.

Table 1.1. Literature review of the skin dielectric properties at 60 GHz

σ Reference T (°C) Method (S/m) In vitro Ghandi and Razi [25] 8.89 – j13.15 43.9 37 ± 0.5 E Gabriel et al. 10.22 – j11.84 39.5 37 E (wet skin) [27] Alabaster et al. [31] 9.9 – j9.0 30 23 M In vivo Gabriel et al. 7.98 – j10.9 36.4 32.5 E (dry skin) [27] Alekseev and Ziskin [33] 8.12 – j11.14 37.2 32.5 ± 0.3 M Chahat et al. [29] 8.02 – j10.5 35 32.5 ± 0.5 M Hwang et al [31] 8.05 – j4.13 24 - 26 M

It has been demonstrated that water [33], sodium [34], and proteins or lipids [35] content of cancer and healthy tissues is different. Typically the total water content of the healthy skin is 65– 75%, while the water content of malignant tissue is higher than 80% [36]. Besides, it has been reported that malignant lesions have higher sodium content than normal skin which determines the retaining of more water [37]. The higher water content of the malignant lesions results in their 18

CHAPTER 1. State of the art

larger values of the complex permittivity as well as of the conductivity compared to normal skin. In turn, this determines significant stronger reflection of the incident wave when impinging on skin tumors compared to healthy skin [36], [37].

Reflection, transmission, and absorption at the air/skin interface The skin electromagnetic properties influence the reflection and transmission at the air/skin interface. The power reflection coefficient decreases from 0.43 to 0.18 as the frequency increases from 30 to 100 GHz [18]. Power transmitted to the body varies significantly depending on the polarization (i.e., parallel or perpendicular) and angle of incidence [38], [39]. Figure 1.4 represents the power reflection and transmission coefficients for both polarizations, at three different frequencies, namely 57 GHz, 60 GHz, and 66 GHz. They have been analytically calculated considering an incident plane wave at the interface of a flat skin models with the electromagnetic properties extrapolated from Gabriel [27]. For normal incidence, 30–40% of the incident power is reflected from the skin; in particular at 60 GHz, roughly 37% of the incident EM power is reflected and 63% penetrate to the body. Note that transmission in the skin is higher for the parallel polarization.

(a)

CHAPTER 1. State of the art

(b) Figure 1.4: Power reflection and transmission coefficients in the 57–66 GHz range: (a) parallel and (b) perpendicular polarizations [38].

The power transmitted to the body decreases exponentially in skin as a function of depth, while peak PD inside skin increases with frequency as absorption becomes more localized. The penetration depth δ (mm) or skin depth, corresponds to the distance a wave must travel before its amplitude has decayed by a factor of 1/e2 i.e., about 37%. It decreases with the frequency (Fig. 1.5). The variation of the power reflection coefficient and penetration depth as a function of the skin dielectric properties at 60 GHz is reported in Table 1.2.

Figure 1.5: Power penetration depth in the human skin. The penetration depth is calculated using permittivity data from [33].

CHAPTER 1. State of the art

Table 1.2. Reflection coefficient (R) and penetration depth (δ) in different skin models

Reference R (%) δ (mm) Ghandi and Razi [25] 41.2 0.43 Gabriel et al. 39.7 0.48 (wet skin) [27] Alabaster et al. [31] 35.7 0.60

Gabriel et al. 37.8 0.48 (dry skin) [27] Alekseev and Ziskin [33] 38.2 0.47 Chahat et al. [29] 37.1 0.49 Hwang et al. [31] 26.45 1.1

Thermal properties of skin To correctly model or predict the temperature elevation induced in the skin by the exposure to MMW is important to characterize its thermal properties such as thermal conductivity k (W/(m·K)) and heat capacity . Thermal conductivity k (W/(m·K)) represents the rate of heat transfer through a unit thickness of the material per unit area per unit temperature difference. It depends on the physical structure of matter and it is a measure of the ability of a material to conduct heat. In solids, interactions between molecules are facilitated by the fact that they are tightly packed, while in liquids and gases intermolecular spacing is much larger and the motion of the molecules is more random, determining a less effective energy transport. For this reason, pure crystals and metals have the highest thermal conductivity values while liquid and gases the lowest. A high value of thermal conductivity indicates that the material is a good heat conductor while a low value indicates that the material is a poor heat conductor or insulator. The thermal conductivity of materials varies with temperature; in gases it increases by increasing the temperature, while opposite behavior is shown by the liquids, with water being a notable exception [40]. The thermal conductivity of certain solids exhibits dramatic increases at temperatures near absolute zero when these solids become superconductors.

Heat capacity (J/K) is an extensive property of matter meaning that is dependent upon the size/mass of the sample. It is defined as the energy required to raise the temperature of a unit mass of a substance by one degree. It is used as a measure of the ability of a material to store thermal energy; it depends on pressure and temperature, or only on temperature for gases. Specific heat capacity (J/(kg·K)) is an intensive variable whose value does not depend on the amount of the 21

CHAPTER 1. State of the art

substance for which it is measured: the heat capacity is divided by the amount of substance, mass, or volume, thus the quantity is independent on the size or extent of the sample. Volumetric heat capacity (J/( ·K)) is obtained by multiplying the specific heat by the mass density ; it represents the ability of a given volume to store thermal energy. Methods for measuring the thermal parameters of skin in vivo are not well established since many factors influence these properties, such as temperature, state of vasodilatation or vasoconstriction, age, gender, and ethnic differences [41]. Thermal properties of skin vary between different layers, and even within the same layer, there exists large non-homogeneity and anisotropy due to the presence of blood vessels involved in skin thermoregulation. Numerous studies in vitro or in vivo performed both on humans and animals are extensively reviewed in [41]. An example of the thermal properties of different skin layers (i.e., epidermis, dermis, and subcutaneous fat layer), typically used to model the heat propagation due to MMW exposure is reported in Table 1.3 [7].

Table 1.3. Skin thermal properties in the epidermis, dermis, and subcutaneous fat layer

k (W/(m·K)) (J/( K)) Epidermis 0.32 4180 Dermis 0.32 3760 Subcutaneous 0.18 1950

1.1.3 General features of interaction of millimeter waves with the human body Shallow penetration depth of the MMW radiation in the skin results in SAR levels that are significantly higher than those obtained at microwaves for identical PD values. This may lead to a significant heating, even for low-power exposures [42]. The SAR or the initial temperature rise rate is inversely proportional to the penetration depth δ [25]. This means that SAR increases very rapidly with decreasing δ or with increasing the exposure frequency. Physically, this can be explained by the concentration of the absorbed energy in a smaller tissue volume. However, it has been demonstrated that in spite of the superficial absorption of the energy, the temperature elevation may appear at regions deeper than the penetration depth, including fat and muscles, due to heat diffusion. In this paragraph, both the electromagnetic (i.e., SAR and PD) and temperature distributions for in vivo model of skin including mono or multi-layered models are described.

SAR and power density distribution in in vivo models The PD distribution in lossy homogenous material at depth z is given by the Beer-Lambert law:

CHAPTER 1. State of the art

(1.7)

2 where PD0 is the incident PD (W/m ) and δ (mm) is the penetration depth. The SAR at a depth z within the skin is given by:

(1.8)

where PD0 ( ) is the incident PD, R is the power reflection coefficient, z (mm) is the depth in the skin, ρ ( ) is the density, and δ (mm) is the penetration depth. Absorption of MMW is directly related to the free water content in the skin, which varies depending on the body part, age, ethnic origin, blood flow, and environmental conditions [7]. In water, δ is between 0.78 and 0.23 mm for frequencies in the range 30–300 GHz [43], meaning that MMW irradiation cannot penetrate deeply into biological objects containing a high concentration of water. Depending on the skin layer, water content ranges from 59 to 72%, determining a non-uniform distribution of the impinging electromagnetic field. The stratum corneum (SC) has the lowest water content (15–40%), and it is mostly in the bound state, hence it is nearly transparent to MMW. The water content of the viable epidermis (i.e., the epidermis minus the SC) is nearly the same as the dermis (70–80%), such that both layers are often considered as a single layer in modelling absorption of MMW. This water distribution within the skin results in the ease MMW energy penetration through the SC, but in the rapid absorption within the deeper epithelium and dermis with no further propagation into the body. The hypodermis is characterized by low water content that is highly variable depending on age and physiological conditions. Accurate determination of the reflection of MMW from skin is necessary to calculate power deposition into the skin and resulting skin heating. Several models have been introduced depending on the number of layers used to represent the skin, including homogeneous or multilayer-models [28]. The homogeneous model does not take into account the different water content of the skin layers (hence the permittivity), therefore, it is not suitable to precisely represent the SAR distribution within the skin. Multi-layer models may consider SC, dermis and epidermis (typically as a unique layer), fat, and/or muscle. However, from the electromagnetic point of view, a model considering only the skin layers (i.e., SC plus dermis and epidermis), is sufficiently accurate to determine the reflection coefficient because of the very small penetration depth of the MMW, which is greatly attenuated by dermis and epidermis. The deeper fat layer and muscles, indeed, have only little effect on the PD and SAR profiles as shown in Fig. 1.6.

CHAPTER 1. State of the art

(a) (b) Figure 1.6: (a) The PD profile for an IPD of 10 mW/ cm2 at 42 and 61 GHz calculated with a multi-layer model including epidermis (E), dermis, divided into papillary (PD) and reticular (RD), and fat. The first thin layer represents the SC. The boundary between E and dermis (at a depth of about 0.1mm) is marked by a wavy line. The arteriolar subpapillary plexus is located at the boundary between the papillary (PD) and reticular (RD) dermis (~0.3 mm) [44]. (b) SAR distribution for an incident PD of 10 mW/ cm2 at 20 GHz in a 3-layer model of tissue including skin, fat, and muscle. Analytical solution (solid line) is consistent with results obtained with numerical calculations. Diagram on the top of the figure shows positions of different tissue layers [7].

In addition to these models, it has been proposed that some of the skin structures, like sweat duct, may act as an imperfect (i.e., non metallic) helical antennas [45]. Specifically, the ducts emerge from the glands that are embedded at the bottom of the dermis and pass through the dermis and epidermis to the pore at the stratum corneum. The tips of the ducts of the eccrine sweat glands have a coil-like structure. Considering the helical nature of the final portion of the duct and the difference in dielectric permittivities between the dermis and the epidermis, it is likely to assume that the ducts could act as low Q helical antenna in the millimetre–submillimetre wave range.

Temperature distribution in in vivo models Electromagnetically induced heating can be evaluated by means of the bio-heat transfer equation (BHTE) introduced by Pennes in 1948 [46]. This equation represents the first quantitative relationship that describes heat transfer in human tissues including the effects of blood flow:

(1.9)

CHAPTER 1. State of the art

where ρ (kg/ ) and c (J/(kg·°C)) stand for density and specific heat respectively, and the subscripts t and b refer, respectively, to tissue and blood; k is thermal conductivity (W/(m·°C)),

(°C) is the local tissue temperature, and (°C) is the arterial blood temperature. The perfusion rate of blood through the tissue is described by (l/s). is the metabolic heat generation, assumed to be homogeneously distributed throughout the tissue of interest as rate of energy deposition per unit volume. represents the volumetric heat input due to the electromagnetic exposure (i.e., MMW). The equation takes into account both the thermal conduction ( )) within the tissue, and the thermal convection ( ) due to the heat exchange between the tissue and the arterial blood. Numerous models, including 1-D monolayer and 1-D multilayer skin models, have been proposed in order to predict the heat transfer due to MMW induced heating within the skin. They are briefly reviewed hereafter.

Homogeneous models The 1-D homogeneous model considers skin as a semi-infinite medium characterized by isotropic thermo-physical parameters, hence the contribution of the reflection and transmission due to the different electromagnetic and thermal properties of the skin layers is neglected. Temperature elevation induced by a normally-incident plane wave at 1 and 10 GHz has been analytically investigated in [47] through the 1-D BHTE that takes into account the effect of surface cooling and blood flow:

(1.10)

where (kg/ ), (J/(kg·°C)), (W/(m·°C)), and (°C) are the density, specific heat, thermal conductivity, and temperature of the tissue, respectively. (°C) is the temperature of the arterial blood entering the tissue. is the product of specific blood flow rate, blood density, and blood specific heat ( and it represents the contribution of the blood convection to the dissipation of heat deposited by the electromagnetic field. Metabolic heat production is not present in the equation since it can be considered independent from short MMW wave exposures [48]. Heat deposition in tissue can be described as follows:

(1.11)

CHAPTER 1. State of the art

where is the incident power density, R the reflection coefficient, δ (m) is the skin depth, and u(t) is the unit step function. Figure 1.7 represents the temperature increase due to a normally-incident plane wave at 60 GHz with PD of 1 mW/ and 5 mW/ , over an average area of 20 , obtained by solving the equation (1.10).

Figure 1.7: Temperature rise for a homogeneous skin model exposed to a plane wave at 60 GHz [42].

Results of this study show that heating due to MMW exposure affects not only the skin but also subcutaneous tissues including fat and muscles. Therefore, in contrast to the electromagnetic dosimetry, for an accurate assessment of thermal effects, a multi-layer model should be used. Moreover, as demonstrated in [49], BHTE allows accurately predictions of local hyperthermia only at low blood flow. On the contrary, the hybrid bioheat equation (HBHE) allows accurate predictions at both low and high perfusion rates by combining the BHTE with the scalar effective thermal conductivity equation (ETCE) adding in the model the blood dependence of the thermal conductivity [49].

Multilayer models Numerous multilayer models able to take into account the anisotropy of the tissues have been introduced to study the temperature elevation induced by MMW exposure at regions deeper than the electromagnetic penetration depth. Notable works have been performed by Alekseev et al. [50], [51], Kanezaki et al. [52], [53], Ziliberti et al. (up to the THz frequency range) [54], [55], Sasaki et al.[56], Ziskin et al. [7], and more recently by Samaras and Kuster [57]. MMW heating has been described in [50] with the use of the HBHE. Four models, represented in Figure 1.8a, have been used to determine the relationship between non-specific tissue blood flow in a homogeneous unilayer model and dermal blood flow in multilayer models. The impact of the 26

CHAPTER 1. State of the art

blood flow on the increase of the induced temperature elevation was evaluated under the exposure to a plane wave impinging ( = 20 mW/cm²) on the skin surface. The cooling effect of blood flow (Fig. 1.8b) and the impact of the thickness of the fat layer (Fig. 1.8c) have been investigated. The greatest cooling effect was obtained in the homogeneous model (M1), while the addition of a thin external layer E (M2) produced only a small reduction of this effect. When epidermis, dermis, fat, and muscle were introduced in the model (M4), the elevations of dermal blood flow producing the same cooling effect were greater than those of non-specific tissue blood flows in M1 or M2. The presence of the fat layer resulted in the appearance of a temperature gradient between the skin and muscle layers. In particular, the greater the thickness of the fat layer, the greater the temperature increment produced by MMW heating (Fig. 1.8c). The parametric study performed by Kanezaki et al. [53] demonstrated that the temperature distribution induced by MMW exposure strongly depends on the geometry and thermal properties of the multilayer model. The authors demonstrated, using the Laplace transform, that the surface temperature elevation in a three-layer model was 1.3– 2.8 times greater than that in the homogeneous model due to the thermally insulating nature of the fat layer. They also showed that the dominant parameter influencing the surface temperature elevation was the heat transfer coefficient between the body surface and air. Zilberti et al. [54], [55] reported on how the uncertainty and variability of the electric and thermal properties of tissues could affect the estimation of temperature elevation produced during MMW exposure using a 1-D three-layer human skin model. Combining all of the relevant parameters (i.e., permittivity, conductivity, heat capacity, tissue thickness, etc.) they showed that the deviation from the reference solution of the maximum temperature elevation in the skin was included in the coverage intervals from –30% to +10% at 100 GHz with a 95% confidence level. Sasaki et al. [56] conducted a Monte Carlo analysis considering both the variation of tissue thickness and the standard deviation of the dielectric properties of the skin measured in vitro to establish the correlation between frequency and surface temperature elevation, as well as frequency and power transmittance into the body in the 10–100 GHz range. A skin-multilayer plane model made by epidermis, dermis, subcutaneous layers, and muscle was used. It was found 1) that tissue thickness significantly affects the variation of both transmittance and temperature elevation, 2) little differences were found among the temperature elevation analysed in different parts of the body, and 3) dielectric properties of adipose tissue do not impact the temperature elevation at frequencies over 30 GHz (in contrast to previous studies). Simplified models for thermal responses of skin and subcutaneous tissues between 6 and 300 GHz have been reviewed by Ziskin et al. [7]. The dependence of tissue heating on anatomical parameters such as skin thickness, or physiological parameters such as blood perfusion, have been analyzed by means of two complementary approaches, including numerical and analytical models. 27

CHAPTER 1. State of the art

The limits of validity of the models and their relevance to setting safety standards are also briefly discussed. More recently, Samaras and Kuster [57] investigated the effect of oblique incidence on the surface of skin. They used a 1-D 4 layers-model including the stratum corneum, viable epidermis and dermis (as a unique layer), hypodermis, and muscle. The reflection coefficient resulting from the exposure to a plane wave of either transverse electric (TE) and transverse magnetic (TM) polarization with a varying incidence angle at the model surface was calculated analytically. Results stress two important points to consider both in numerical and experimental exposure assessments and in development of exposure guidelines/standards: 1) the importance of the introduction of the stratum corneum to make more realistic modelling as it increases the power transmittance, and 2) that evaluating only the normal incidence may determine an underestimation of the exposure by more than a factor of 10. A new methodology to predict and analyse thermal distribution, thermal-induced mechanical deformation, and thermal-mechanical damage of soft tissues during thermal ablation is proposed in [58] by combining non-Fourier bio-heat transfer, continuum mechanics, as well as non-rigid motion. This study demonstrated that although the Fourier (e.g., it assumes infinite speed of heat propagation in soft tissues) and non-Fourier models have similar accuracy for temperature prediction after a sufficient time of heating, the non-Fourier bio-heat model accounts for the transition state of bio-heat transfer, appearing more suitable to predict the tissue’s instantaneous thermal-mechanical behavior.

(b)

CHAPTER 1. State of the art

(a) (c) Figure 1.8: (a) Tissue models used for modeling MMW wave heating. In model 1, the skin and underlying tissues are considered as a homogeneous tissue. In models 2 and 4, E stands for the epidermis. In model 4, D stands for the dermis. The deepest layer in each model has semi-infinite thickness. (b) Temperature elevation on the skin surface as function of the blood flow. (c) Temperature elevation within tissue calculated for different thicknesses of the fat layer (2, 4, and 8 mm) in model M4 [50].

1.1.4 Tissue-equivalent models Experimental tissue-equivalent phantoms have been extensively used in research and compliance testing to quantify the exposure levels induced in the human body by a portable wireless device or to investigate the electromagnetic propagation around and inside the human body [59]. The most common classification is based on the physical state of the phantom, namely liquid, semisolid, and solid. Liquid phantoms consist of a container, typically mimicking the shape of the body’s part of interest, filled with a liquid having the average tissue-equivalent dielectric properties in the considered frequency range. These phantoms are extensively used for the conservative assessment of the SAR in the frequency range from 30 MHz to 6 GHz [60], but they have a limited interest at MMW. The presence of the shell, indeed, does not permit measurements at the phantom surface where most of the energy is absorbed. Semisolid phantoms are usually made by water-based materials capable of self-shaping using a solidifying agent (i.e., agar). Therefore they do not require any bounding container. The required electromagnetic properties are adjusted by varying the percentage of the materials. For instance, the first phantoms simulating both high (e.g., muscle or skin) and low (e.g., fat or bones) water content tissues have been developed by Guy et al. [61] in the 0.2–1 GHz range. High water content phantoms were composed by water, sodium chloride, TX–150, and polyethylene powder. Typically, the relative permittivity can be varied over a wide range by varying the percentage of polyethylene powder, while the electrical conductivity can be modelled by varying the salinity of the material. 29

CHAPTER 1. State of the art

Low water content phantoms were realized by using laminac polyester resin, catalyst, acetylene black (to adjust the conductivity), and aluminium powder (to adjust the dielectric constant). Today, numerous phantoms are available to represent several parts of the body in different regions of the electromagnetic spectrum: 1) muscles and brain at 0.3–2.45 GHz [62], 2) 2/3 muscle-equivalent phantom at 3.1–10.6 GHz [63], and 3) skin at 10 GHz [64]. At higher frequencies (i.e., MMW range) several phantoms are currently available mimicking the dielectric properties of skin [29], [65]–[67], muscle [65], or breast [68]. The first broadband (55–65 GHz) tissue equivalent phantom in the MMW range has been realized by Chahat et al. [29]. The main constituents are 1) deionised water that primarily determines the dispersive behaviour of the phantom, 2) agar for the retention of self-shaping (its contribution to the phantom dielectric properties is negligible for small concentration typically below 4%), 3) polyethylene powder to decrease the real and imaginary parts of the permittivity, 4) TX–151 to increase the viscosity of the agar solution allowing the agar and polyethylene powder to mix, and 5) sodium azide used for the phantom preservation. The fabrication procedure is detailed in [69]. The thermal properties of this phantom, such as specific heat capacity and thermal conductivity, were experimentally characterized in [70] using a differential scanning calorimeter and thermal method of guarded hot plate, respectively. The thermal behaviour of the phantom was studied for several phantom thicknesses (5, 10, and 15 mm) analytically by solving the 1-D heat transfer equation, and numerically through electromagnetic-thermal co-simulations. Results were finally validated by measurements using an experimental setup based on high-resolution infrared thermometry. Results obtained are important for temperature-based dosimetric assessment of human body heating during MMW exposure. Moreover, they cannot be directly used for predicting temperature increase in skin. In [65], muscle and skin-equivalent phantoms, for the 26.5–40 GHz range, with a similar composition as the one in [69] have been proposed. The main constituents of the phantoms are deionised water, agar-agar, guar gum, polyethylene, and gelatin powder. Water and agar-agar are common to both the models; polyethylene powder and guar gum (a low-cost natural polysaccharide which is used as an emulsifier equivalent to TX–151) were employed for the realization of the muscle-equivalent phantom, while gelatin powder was used for the skin- equivalent phantom. The validity of the skin-equivalent model proposed in [65] has been extended in the V and W bands (up to 100 GHz) by the group of Lacik et al. in 2016 [67]. More recently, a novel composition of a phantom mimicking the dielectric properties of skin in the frequency band within 24.25 and 67 GHz has been analysed [66]. Several combinations of water and three different alcohols, namely ethanol, methanol, and 1-propanol at different concentrations (i.e., 5%, 10%, and 15%) were tested. The solutions were gelled by using k-carrageenan, which has promising 30

CHAPTER 1. State of the art

mechanical properties without altering the dielectric ones. The addition of 1% of CaCl2 was necessary to allow its dissolution into water. A breast-equivalent phantom has been proposed by the group of Di Meo et al. [68] in the 0.5–50 GHz frequency band. Water, sunflower oil, emulsifier (i.e., dishwashing liquid), and two different solidified agents (i.e., gelatin and oil-waste hardener) have been employed. The accurate mimicking of both healthy and cancerous tissues, at different adipose content, have been obtained by varying the concentration of oil. Tissues with a significantly low water content (i.e., highly fatty) were realized by using the oil-waste hardener that allowed the mixing of high concentration of oil and water, otherwise not possible with gelatin. Semisolid phantoms are low-cost, easy to fabricate and to assemble into the desired shape or in several different layers mimicking the anisotropy of the body tissues. The main drawbacks of these phantoms are 1) the deterioration over time due to the water evaporation, resulting in the degradation of the dielectric properties, and 2) the temperature dependence of the electromagnetic properties because of the high water concentration.

(a) (b) (c) Figure 1.9: (a) Skin and muscles-equivalent phantoms in the 26.5–40 GHz [65], (b) skin-equivalent phantom representing an arm and a hand in the 55–65 GHz [29], and (c) breast equivalent model with tumor inclusions [68].

Solid phantoms are mainly used for body surface SAR measurements, antenna measurements, and for studies of propagation around or inside the body. They are generally based on 1) a mixture of ceramic and graphite powder [71], [72], 2) a combination of carbon fibers and silicone rubber [73], and 3) conductive plastic containing carbon black [74]. These phantoms can be designed to fit a wide range of complex permittivities up to 60 GHz [75] and have excellent mechanical and dielectric properties that do not degrade over long time. However, the realization of solid phantoms at MMW, mimicking the skin electromagnetic properties, is very challenging due to the high loss tangent of skin (tanδ ≈ 1.3 at 60 GHz). For certain applications, such as study of impact of the presence of the human body on the antenna performances or characterization of the body-centric 31

CHAPTER 1. State of the art

propagation channel, it is sufficient to reproduce the same power reflection coefficient as that of skin without necessarily having the same electromagnetic properties. In [75], the first solid skin- equivalent phantom to emulate the same reflection coefficient at the air/phantom interface as at the air/skin interface has been designed for the 58–63 GHz range. It consists of a lossy flexible PDMS (silicon-based organic polymer) dielectric sheet containing carbon powder with a metallic backing (Figure 1.10a). The phantom is lightweight, easy to fabricate, and preserves its dielectric properties for a long period of time since it does not contain water. The main drawbacks of such phantom are the cost and the requirement of specific equipment and production conditions for fabrication, including high temperature and high-pressure manufacturing procedures. Recently, a new solid phantom known as POPEYE10-3TO110 (POsable Phantom for Electromagnetic sYstems Evaluations 10 [76]) has been manufactured by SPEAG (Fig. 10b) for the 3–110 GHz frequency band [76]. The phantom consists of a) a torso phantom (TORSO-OTA-V5.1), b) posable right and left arms (ARMR/L-V5.5 or ARM3TO6R/L-V5.5), c) posable right and left legs (LEGR/L-V5.5 or LEG3TO6R/L-V5.5), d) buttocks phantom (BUTT-V5.5 or BUTT3TO6-V5.5), and e) right and left feet (JIAOR/L-V5.5 or JIAO3TO6R/L-V5.5). Except for the torso shell, manufactured from a vinylester-reinforced fiberglass low-loss structure, the rest of the phantom is made by lossy silicone-carbon-based mixture. A special low-loss silicone coating is applied to POPEYE10- 3TO110 to extend the frequency range to 110GHz.

(a) (b) Figure 1.10: (a) PDMS–carbon composite with a metallic backing [75]. (b) POPEYE10-3TO110 [76].

CHAPTER 1. State of the art

1.2 In vitro studies at millimeter waves

In vitro studies are performed to elucidate the mechanisms of interaction between MMW and living tissues at cellular and molecular level. During the exposure, cells are seeded in plastic vessels (i.e., test tubes, tissue culture plates [TCPs], flasks, or Petri dishes) which dimensions vary depending on the particular requirement of the experiment. Accurate electromagnetic and thermal analysis of the power-dependent threshold and potential combined electromagnetic/thermal induced effects are of utmost importance for the correct interpretation of the biological results. This requires the development of specific exposure systems [77] providing the control and the characterization of the exposure level within the samples, and the optimization of the exposure conditions (uniformity of the field distribution, number of simultaneously exposed samples, etc.). The greater the accuracy of the dosimetry, in terms of characterization of the electromagnetic field, SAR, and temperature distribution within the exposed biosample, the greater the reliability of the biological and biophysical data obtained experimentally. In this section the exposure systems employed for biological studies at MMWs are firstly introduced. Successively, both numerical and experimental dosimetry are described. Finally, a detailed description of the devices used to perform measurements of the MMW-induced temperature elevation in the sample under test (SUT) is provided.

1.2.1 Exposure systems Specific exposure systems have been used to experimentally investigate the possible biological impact of MMW exposures. These systems are generally made up by four subunits [39] detailed hereafter. Signal generation sub-unit. It consists of electronic components, such as generator, amplifier, attenuator, isolator, and radiating structure/antennas. Spectrum and power control sub-unit. It consists of a VNA (Vector Network Analyzer) or power meter to ensure real-time monitoring of the signal spectrum and output power stability necessary for the reliable interpretation of the biological results. Exposure chamber. It ensures the appropriate environmental conditions for the specific biological SUT and determines the electromagnetic boundary conditions. For instance, for cells in culture, precise temperature control (37 °C) and pH medium (pH = 7.5) must be ensured.

CHAPTER 1. State of the art

Temperature interface. It ensures the control of the temperature elevation induced by the electromagnetic exposure. Several devices are suitable for the monitoring of the MMW-induced heating (section 1.2.5, Technique for measurements of MMW-induced heating). In most of the reported in vitro studies, far-field exposure set-ups using a directive antenna, are employed for the exposure of biological samples under free-space conditions [78]–[80], or within a thermo-controlled incubator [81]–[84]. Far field exposure set-ups allow simultaneous exposure of several samples, and provide a nearly uniform distribution of the PD over a sample set (typical deviations of the PD are less than 30%) as well as a low peak-to-mean power density ratio. However, the efficiency of far-field exposure is low, resulting in low PD levels reached by the samples. Higher power density can be achieved either increasing the output power of the generator [85], or putting the samples closer to the antenna by characterizing the near-field distribution of the biological samples in the reactive zone [86]–[88]. The latter solution requires an accurate dosimetry and optimization of the near-field exposure as significant deviations of the SAR may appear even for small displacements (several millimeters) between the antenna and the sample. Uniformity of the exposure is required to assure that all the cells receive the same amount of irradiation dose. However, this condition is difficult to achieve due to the very short wavelength of the MMW which causes a highly irregular irradiation path [89]. Different solutions have been proposed to improve the uniformity of the irradiation pattern by 1) enlarging the culture dish diameter or reducing the culture dish bottom thickness [90], or 2) irradiating the SUT with a novel choke ring antenna (CRA) [91]. The latter has been optimized to enhance i) the uniformity of the PD distribution at the surface of the SUT (i.e., by a factor of 1.5 to 2 compared to standard waveguide feed), and ii) to increase the incident PD.

1.2.2 Biological effects of millimeter waves A large amount of in vitro research has been performed to figure out cellular and molecular mechanisms of possible MMW interactions, although results often suffer of poor reproducibility and reliability. Principal studies have been collected in four extensive reviews by Packhomov et al. [92], Ramundo-Orlando et al. [93], Le Dréan et al. [94], and more recently by Alekseev and Ziskin [18] up until 1998, 2010, 2013, and 2018, respectively. The bio-physical mechanisms underlying the MMW actions are still poorly understood. It has been demonstrated that at 10–100 GHz, heating is the major effect arising from the motion of water molecules following the absorption of the MMW energy at any intensity [18]. Depending on the IPD of the exposure, MMW induced effects can be divided into two categories, namely indirect

CHAPTER 1. State of the art

thermal effects, mediated by the significant temperature elevation of the exposed biological structures to IPD > 5 mW/cm2, and direct non-thermal (or athermal) effects to low IPD < 5 mW/cm2 where only an insignificant temperature elevation < 0.1°C is observed [95]. High-intensity MMW may act in a dose-dependent manner: significant thermal responses (ΔT > 0.5°C) appear after exposure to IPD > 5 mW/cm2 , and in case of very high-power exposures (above several hundreds of mW/cm2), it can lead to a pain sensation or tissue damage [39]. Rapid temperature increase during irradiation, even if only by a small total increment (i.e., 0.0025 °C/s), can cause certain biological effects such as changes in the firing rate of neurons [96], [97], or increase of both neurite outgrowth [98] and microtubule assembly rate [99]. The existence of specific electromagnetic effects is highly controversial. The main findings reported at cellular and molecular level are summarized hereafter, including the effects on i) cell cycle and proliferation, 2) gene expression, and 3) biological membranes. Several studies performed in vitro demonstrated that the exposure of cell cultures to MMW may have beneficial consequences in cancer treatments due to the direct antiproliferative effect of this radiation. Few studies from the group of Beneduci et al. showed that that low power (< 50 μW) MMW at 53–80 GHz may cause selective inhibition of tumor cell growth and morphological alterations in human malignant melanoma [100] or breast cancer cell lines [101]. However, results are very contradictories; for example, another study, provided by the same group [102] demonstrated that the 42.20 and 53.57 GHz do not affect the proliferation rate or the cell cycle of melanoma cells. Moreover, cellular growth and viability of immortalized keratinocytes cells derived from human epidermis is not affected by 60 GHz at IPD lower than 1 mW/cm2 [81]. Li et al. [103] observed that the exposure to MMW at an intensity of 4 mW/cm2 caused progression of the cell cycle of in vitro of chondrocytes promoting their proliferation. Recently, Haas et al. [98] demonstrated that 24 h exposure to MMW at 60.4 GHz (IPD = 10 mW/cm2) induces a slight increase of neurite outgrowth in neuron-like cell line although the change is related to thermal effects (i.e., temperature increase ≈1–2 °C). The study of genetic expression represents a fundamental step to deeply and quantitatively evaluate the behaviour of cells following the MMW irradiation. Therefore, a large number of studies has been performed especially around the 60 GHz band. MMW are nonionizing radiations and, as expected, are not genotoxic, e.g., they do not damage the DNA [104], [105]. On the contrary, proteins are particularly fragile and possibly affected by relatively weak external stresses (e.g., heat), arising the hypothesis that proteotoxic stress may be triggered by MMW exposure. The potential MMW-induced proteotoxic effects have been assessed mainly by analysing 1) the expression of chaperones molecules [79], [106] (i.e., the natural cellular response for the protection from protein unfolding and aggregation [107]), 2) the expression of mRNA [82], [108], and 3) the 35

CHAPTER 1. State of the art

stress induced on the endoplasmic reticulum (ER) [109], [110]. Low power MMW energy, which does not trigger significant temperature elevation, do not modify the stress-sensitive gene expression of chaperone proteins such as HSP70 and Clusterin [83], [106], nor is able to trigger the ER stress [109]. DNA microarray represents the most powerful tool to profile the global RNA content, therefore, using a microarray-based approach, Le Quement et al. [82] analysed modifications of the whole genome of a human keratinocyte model exposed for 1, 6, or 24 h to 60.4 GHz with an average incident power density of 1.8 mW/ cm2 and an average specific absorption rate of 42.4 W/kg. High-throughput DNA microarray analyse evidenced 120 genes differentially expressed after 6 h of exposure on 26300 expressed genetic entities. However, only 5 genes were confirmed as really statistically differentially expressed by using quantitative RT-PCR, showing that impact of MMW exposure is extremely modest. The effects of MMW on cell metabolism were tested by co-treating MMW exposed cells with a glycolysis inhibitor, 2-deoxyglucose (2dG) that decreases ATP content [111]. Whole genome expression was evaluated and no change in gene expression was found when cells were exposed to MMW alone under athermic condition. The MMW/2dG co-treatment slightly altered the transcription of the 2dG-sensitive genes (6 genes on 632 were found to be differentially expressed compare to 2dG treatment), indicating the capacity of MMW to possibly interfere with the bioenergetic stress response. Pure electromagnetic (i.e., non-thermal) effects are mainly attributed to the interaction of the MMW radiations with the cellular membranes. Theoretical studies by Frölich [112], [113] seem to justify this hypothesis suggesting that membrane are sensible to coherent excitations above 109 Hz. Nevertheless, this has never been proved experimentally thus far. Cell membrane consists of phospholipid bilayer with embedded proteins; numerous artificial lipid membranes have been largely used in experiments as they are simple and convenient models for the study of MMW effects on ion transport, phase transition, hydration membrane surface charge, and lipid peroxidation. Different effects have been observed including: 1) elongation, induced diffusion of fluorescent dye di-8-ANEPPS, and increased attractions between giant phospholipid vesicles exposed to 53.37 GHz at 0.1 mW/cm2 [114], 2) significant structural changes at the water-bilayer interface for exposures between 53–78 GHz [115], 3) increase of lateral pressure of phospholipid monolayer at 60 GHz and PD = 0.9 mW/cm2 [84], 4) modifications in ion transport through the bio- membranes [116], and 5) changes in Ca+ regulation [117]. Studies performed on in vitro cells exposed to MMW at IPD of 34.5 mW/cm2 revealed transient externalisation of phosphatidylserine (PS), a lipid normally located in inner-leaflet of cell membranes [118]. However, some of the studies performed suggested that MMW exposure produced significant thermally mediated effects on excitable cells via basic thermodynamic mechanism. For example, Alekseev and Ziskin [96] 36

CHAPTER 1. State of the art

demonstrated that the exposure to MMW at 60.22–62.22 and 75 GHz on A-type K+ and Ca + currents of Lymnaea neurons resulted in the increase of the peak amplitude, activation and inactivation rates of both ionic currents mediated by the temperature rise of the irradiation. Recently Shapiro et al. [119] analysed the effect of 60 GHz radiation on the voltage-gated sodium potassium channels in oocytes. The observed effects (i.e., changes in the kinetics and activity levels of voltage-gated potassium and sodium channels and a sodium-potassium pump) were consistent with a thermal mechanism. The analysis of the main biological effects reported in vitro studies revealed numerous contradictions suggesting that studies are still needed to clarify both thermal and non-thermal effects of the MMW radiations under well- controlled and reproducible conditions.

1.2.3 Numerical dosimetry Numerical dosimetry is fundamental to characterize the exposure from the electromagnetic (i.e., SAR distribution), thermal (i.e., temperature distribution), and thermodynamic (i.e., development of convection) points of view. The major challenges for numerical dosimetry at MMW are: 1) electrically large problems ( varies from 2.5 to 1.25 mm in 30–100 GHz range [39]; this implies small mesh cell sizes of the numerical models of the order of 0.1 mm [120]), 2) uncertainties on the precise values of the dielectric properties of tissues and absence of well- established database beyond 20 GHz (section 1.1.2, Electromagnetic properties), and 3) multi-scale problems related to the presence of the electrically small sub-structures (e.g., cell monolayers or different layers of skin [120]). The electromagnetic problem must be coupled with the thermodynamic one to carefully take into account possible heating and dielectric constant variations related to thermal gradients. In turn, high temperature gradients within the culture medium used to provide nutrients to the exposed cells, originate convective currents which may perturb the concentration of the substances carried by the medium or provide an additional movement to the cell monolayer, which is absent in the sham sample used as negative control [121]. In the MMW range the main contributions to numerical dosimetry for in vitro experiments came from the works of Zhao et al. (i.e., [89], [90], [122], [123]), and Zhadobov et al. (i.e., [81], [86], [87], [120]). Despite the studies provide rigorous characterization of the electromagnetic field or SAR distribution (in the culture medium and/or in the cell monolayer) by means of different numerical methods (i.e., finite-element method (FEM), finite integration technique (FIT), or finite- difference time-domain (FDTD)), calculated temperature distribution is not always provided and

CHAPTER 1. State of the art

thermodynamic events are typically neglected. If provided, experimental measurements are performed to determine temperature distribution in the exposed SUTs such as in [86], [124].

1.2.4 Experimental dosimetry Experimental dosimetry is typically performed through: 1) E-field-based probe systems [125], and 2) temperature measurements on the surface or inside the exposed tissue or phantom. However, at MMW, the direct field-based dosimetry faces three major problems. First, highly localized absorption of the MMW within the biological tissues and solutions determine high gradients of PD and SAR. This implies that the measurements devices must have spatial resolution better than 0.1– 0.2 mm. Second, the already-existing E-field probes are too big in size for the local dosimetry, and additionally may perturb the electromagnetic field and temperature distribution. Third, the sample conductivity should be known to determine the local or average SAR. Therefore, this experimental technique has a limited practical interest. In alternative, the most efficient way to experimentally determine the SAR and the temperature distribution consists in remotely or invasively measuring the near-surface thermal dynamic of the exposed SUT [126].

1.2.5 Techniques for measurement of MMW-induced heating Measurements of the MMW-induced temperature elevation play a fundamental role for 1) the quantification of the temperature distribution in the exposed SUTs, and 2) the retrieval of the SAR. It has been demonstrated that the initial phase of heating kinetics is proportional to SAR and can be determined as follows [126], [127]:

(1.12)

where (J/(kg K)) is the specific heat, dT (°C) is the temperature rise, and dt (s) is the heating time. The most accurate way to determine is to fit the experimental heating kinetics to an appropriate theoretical thermal model as detailed in next chapter (paragraph 2.7, Temperature- based retrieval of SAR). Measurement systems technologies for thermal dosimetry at MMW can be classified into two groups, namely invasive and non-invasive. Invasive techniques necessitate the insertion of the instrument into the sample where temperature has to be measured; on the contrary non-invasive instrumentation permits the measurement from the outside of the sample without disturbing its structure. Point, 2–D, and 3–D temperature elevation within the exposed sample can be obtained 38

CHAPTER 1. State of the art

depending on the characteristics of the specific device used. Optical fiber (OF) thermometers and micro-thermocouples (TC) are invasive instruments that give information about the temperature rise in a specific point of the exposed SUT. Infrared (IR) cameras or thermochromic liquid crystals (TLCs) allow to obtain 2–D temperature distribution in a non-invasive or invasive manner, respectively. Magnetic resonance thermal imaging (MRTI) non-invasively maps the 3–D temperature distribution of the exposed object. OF temperature sensors work by modulating one or more properties of a propagating light wave including intensity, phase, polarization, and frequency in response to a temperature change. Extrinsic optical sensors use the fiber only as a mechanism to transmit light to and from a sensing element, while intrinsic optical sensors use the optical fiber itself as the sensing element [128]. An optical fiber is made up by a cylindrical wire in plastic or glass, the core, used to transmit the light, surrounded by a cladding, and coated by a protective layer known as buffer. A schematic of the basic structure of the OF sensor is given in Figure 1.11. It consists of an optical source (laser, LED, laser diode, etc.) connected to an optical fiber through which light propagates, ending with the sensor that transduces the temperature into an optical signal which is successively transferred to the detector, and finally to the processing electronic block. A large number of OF thermometers are commercially available, and depending on the nature of the sensor and the type of modulation exploited, they may be classified in [129]: i) fluorescence decay, ii) semi-conductor crystal (typically gallium arsenide (GaAs) band-gap variation, iii) Fabry-Pérot interferometer, and iv) Fiber Bragg grating sensor. The use of OF thermometers is particularly suitable for temperature measurements in strong electromagnetically influenced environment due to the fact that the probe is completely non-metallic and the sensor is nonconductive offering complete immunity to the electromagnetic field. Besides, due also to their reduced size and cost, low weight, small dimensions, and geometrical versatility, they are extensively used for the accurate temperature measurements during minimally invasive thermal treatments [130] of cancer. Due to their chemical and thermal resistance they can be used for environments too harsh for electronic sensors. They allow measurements in a large range of temperatures (typically several hundred of degrees), with a precision of the order of ± 0.1 °C. However, they are invasive instruments that need to be inserted into the SUT to obtain the measure, and moreover their spatial resolution is limited (only point temperature can be obtained). OF sensors have been largely used during in vitro studies at MMW [86], [108], [131], [132]. Recently, it has been demonstrated [132] that FO probes fail to adequately measure local heating induced by MMW exposure, and in particular, short temperature pulses with a duration of 1 s, because of the great time constant (0.6 s) of their input.

CHAPTER 1. State of the art

Figure 1.11: Basic components of an optical fiber sensor system.

A TC is a temperature-voltage transducer, which is made by welding two different metals at one end. Commercial TC are inexpensive, simple to use and able to measure a wide range of temperatures (i.e., for a K-type TC, –200 to 1200 °C). They allow for accurate local measurements of temperature with a time constant of the order of µs, which is desired when measuring rapid temperature variation (i.e., sharp MMW-induced thermal pulses with duration of several seconds [132]). The main limitations are i) the accuracy (systematic errors of less than one degree °C can be difficult to achieve), ii) the invasiveness, and iii) possibly induction of currents. However, they represent the optimal device to measure the temperature elevation in the cell monolayer, during in vitro experiments at MMW by providing that: 1) the thickness of the thermocouple is comparable with the thickness of cell monolayer (< 1mm) [133], and 2) the lead of the TC is oriented perpendicular to the E-field [127]. Several studies have exploited TC to accurately determine the local temperature elevation induced by MMW exposure during in vitro studies such as [96], [118], [132].

CHAPTER 1. State of the art

Figure 1.12: Temperature dynamics measured by means of micro-TCs during in vitro experiments. Curve 1 and 2 refer to measurements performed with a 25 µm and 75 µm TC, respectively [132].

IR camera is a non-contact device that detects infrared radiation and converts it into the temperature map on the surface of the object under test [134]. The systems are very fast, almost instantaneous, and they are contactless, therefore they are not distorted by the material of the target, nor is there any energy loss, transfer, interaction, or contamination of the SUTs. The main drawback is related to the fact that IR devices are limited to surface measurements. Therefore, if they appear perfectly suitable instruments for superficial MMW heating [29], [49], [135], they are not eligible to perform in vitro measurements of the temperature elevation induced inside the exposed medium and/or cell monolayer. In fact, irradiation of the cell monolayer is typically performed from the bottom of the TCP due to the shallow penetration of the MMW. This means that measurement from the top of a monolayer is impossible due to the strong absorption of radiation by the medium, while measurement from the bottom of the well is complicated by the presence of plastic between the cell monolayer and IR camera.

Figure 1.13: IR image of the temperature distribution on the surface of phantom exposed at 60.4 GHz by an open-ended waveguide [30].

The term liquid crystal refers to a mesomorphic phase between the crystalline solid state and the isotropic liquid state, sharing specific properties of both states. For example, they can be treated as liquids with regard to their mechanical behavior (i.e., fluidity or surface tension), but also as solid with regard to their optical properties (i.e., anisotropy to light or birefringence) and molecular orientation. The word thermochromic associated to liquid crystals refers to their ability to change their color as a function of the temperature. TLCs are optically active (they can rotate the plane of linearly polarized light) nematic liquid crystals characterized by chiral (twisted) molecular structures [136]. The chiral nematic mesophase is arranged in layers, oriented as a helix and in each layer molecules are ‘rod-shaped’ (Fig. 1.14a). The distance for a full twist in the helix is called pitch

CHAPTER 1. State of the art

(p) and it determines the wavelength of light which is reflected (Bragg diffraction). As the temperature changes, the pitch of the liquid crystal changes, determining a variation of scattered light which is proportional to the temperature surrounding the TLC (Fig. 1.14b). Hence, the color reflected by TLC can be directly related to temperature [137], [138]. TLC may turn from colourless (black) to red at a given temperature, and as the temperature increases, they pass through the other colours of the visible spectrum in sequence (orange, yellow, green, blue, violet) before turning colourless again at higher temperatures. The difference between the red-start and the blue-start temperatures is defined bandwidth and it can be customized depending on the specific application. Since the introduction of TLC for thermal mapping in 1963 [137], a large amount of work has been performed spanning from the study of the gas turbine heat transfer, flow visualization, microfluidic devices, and medical diagnostics [139]–[141]. Nowadays, different TLC mixtures are commercially available [142], including i) unsealed TLC used as particle tracers in liquid, ii) microencapsulated (slurries and coatings), which are more stable though less bright than unsealed particles, and iii) coated printed sheets. The working temperature interval ranges from –30 °C to 115°C as reported in the handbook of thermochromic liquid crystal technology provided by Hallcrest [136]. Images of the color play are captured by commercial imaging hardware such as digital video cameras, and processed after the experiment using image-processing algorithms to yield full-field temperature measurements. Calibration of the acquisition system is of outmost importance to obtain reliable quantitative data of the temperature map [141]. Applications in the MMW frequency band include few studies performed on rabbit eyes by the groups of Kojima et al. [143]–[145]. For example, microencapsulated TLC (0.2%) have been injected in the rabbit eyes to observe the temperature elevation and the development of convective currents after the exposure to 200 at 18 GHz and 40 GHz, as shown in Figure 1.14c [143]. At the best of our knowledge, the possibility to use TLCs for the temperature monitoring during in vitro studies at MMW has never been considered.

(b) 42

CHAPTER 1. State of the art

(a) (c) Figure 1.14: (a) Model of TLC organized in layers where rod-shaped molecules lay (zoom). The spacing p (pitch) between the layers corresponds to the rotation of 360°of the long axes [139]. (b) Working principle of the TLC based on the Bragg diffraction. (c) Temperature distribution in the anterior chamber of the rabbit eye measured through TLC at 18 GHz and 40 GHz (200 after 5 s (A,D), 10 s (B,E) and 30 s (C,F) following the start of the exposure [143].

MRI exploits the magnetic properties of the protons (i.e., the hydrogen nucleus) present in the water molecules that compose the majority of the body tissues. When a strong magnetic field is applied, the water protons align themselves with the field, and when the field is turned off, they return in the initial position. The contrast between different tissues is determined by the rate at which excited atoms return to the equilibrium state [146]. MRTI may provide the 3–D mapping of the temperature distribution of the SUTs by exploiting the temperature dependence of the water proton resonance frequency (PRF) [147]. Recently MRTI scanning have been used to non- invasively quantify 3–D temperature elevation and SAR variations induced by MMW [148]. This method provides numerous advantages: 1) it is sensitive to small temperature changes (< 0.1°C) allowing SAR calculation with millimeter resolution [150], 2) it provides volumetric thermometry and SAR information of the entire phantom in about 1 min [134], and 3) it allows local energy measurements at the interface of the exposed sample (where the MMW energy is mainly focused). SAR reconstruction from high resolution 3–D MRTI temperature mapping can be obtained by using the heat equation inversion (HEI) which includes terms relative to heat diffusion and conduction effects [151]. Experimental and numerical results provided in [151] are compared in Figure 1.15 demonstrating the validity of the methodology. At the best of our knowledge, the technique has

CHAPTER 1. State of the art

never been applied for the measurement of the temperature elevation during in vitro experiments in the MMW frequency band.

Figure 1.15: Simulated temperature change and 1 and 10 g average SAR from a simulated dipole (top row). Experimental temperature measurement using MRTI and reconstructed 1 and 10 g average SAR maps using HEI method (bottom row) [151].

1.3 Heat-induced biological effects: from continuous to pulsed- modulated waveforms

Electromagnetically induced heating has been largely used as a means of treating numerous types of cancer [152]. The following section will firstly provide an historical overview of the evolution of the thermal treatments; successively, the thermal therapies currently used in the practical clinic are described, and the main physiological and cellular effects induced by heat are presented. Afterwards, the concept of thermal dose is provided, and finally the importance of exploiting pulsed heating for therapy optimization is detailed.

1.3.1 Thermal treatments

History of thermal therapies for cancer treatment

CHAPTER 1. State of the art

Therapeutic applications of heat date back to the ancient Egypt around 2600 BC. According to the Edwin Smith Papyrus [153], a record of scientific approaches to medicine in Ancient Egypt, Imhotep an Egyptian polymath, burned off masses growing on the breast with a heated poker. Successively, by 2000 BC, local destruction of tumors became widely used [154]. In the ancient Greece, Hippocrates successfully used heat to treat breast tumors [155]. The first paper on hyperthermia was published in 1886 by Busch [156] who reported that a malignant sarcoma underwent to a complete remission when a fever was caused by erysipelas infection. On the basis of this funding, the american surgeon Coley in the same years, extracted toxins from Streptococcus pyogenes and Serratia marcescens and used them to induce fever for treating patients with cancer [157]. Despite promising observations, during several decades, these cancer treatments were difficult to administer in a controlled manner, and responses were unpredictable [158]. Further investigations have been performed throughout the first half of the 20th century [159] but the use of thermal therapies was not recognized as a separate treatment method until the 50s, and it existed predominantly in the form of febrile therapy, where the thermal effect was a part of the complex body reaction. The modern discipline of thermal therapies have risen from the studies performed between the 1960–1980 which demonstrated highly quantifiable, time dependent cytotoxic effects of heat on cells [160]. A turning point is constituted by the studies of Crile in the 1960s [161]– [163]. He demonstrated by using transplanted tumors in mice new biological phenomena: i) heat cytotoxicity was dependent on time and temperature, ii) increased sensitivity of large versus small tumors to heat (later attributed to vascular events), iii) heat-induced thermotolerance (i.e., cellular resistance induced by heat) of normal and tumor tissues, and iv) heat-induced sensitization to radiation. These promising observations frame the rationale for the clinical use of hyperthermia and the development of more effective technologies for the precise application of heat to tumors as well as for the control of the temperature distribution during the treatment. In the past 20 years the development of the heating methods and the improvement of the techniques to provide real-time control of the temperature within the tumor (i.e., non-invasive techniques using MRI or Computed Tomography (CT)) have transformed thermal therapies into a promising treatment for many patients, especially those with surgery intolerance [152]. Nowadays, thermal therapies can be classified into hyperthermia and thermal ablation depending on the temperature range attended and the duration of the treatment.

Hyperthermia Hyperthermia relies on the use of temperatures between 40 and 47°C. Depending on thermal dose, thermal treatments may result in mild hyperthermia (40–41 °C for 6–72 h) or moderate- 45

CHAPTER 1. State of the art

temperature hyperthermia (42–47 °C for 15–60 min) [155]. Temperature elevation within the 40– 47°C temperature range sensitizes cells to other treatment modalities, and augments the efficacy of radio and chemotherapy with insignificant or no injury to normal tissues as detailed below (section, Cellular and molecular responses to heat). For this reason, hyperthermia is typically applied in combination with other treatments, such as chemotherapy, radiotherapy, surgical treatments, gene or immunotherapy, and rarely alone. Hyperthermia can be used with all stages of cancer, although its current main use is with advanced solid tumors that are hardly operable or inoperable, and with recurrent tumors or metastases. There are three main clinical methods currently employed, namely local, regional, and whole body hyperthermia, depending on the organ to be targeted, the stage of the cancer, and the energy distribution technique [164]–[166]. Local hyperthermia permits the treatment of relatively small tumors (≤ 3 cm up to 5–6 cm), located superficially or within an available body cavity such as the rectum or esophagus. The heating source can be either external or internal. Regional hyperthermia targets much wider regions than local hyperthermia, such as an organ or a region of the body. It is mostly used to treat deep tumors in the pelvis or abdomen. This method is more complex than local hyperthermia because of the complexity and non-homogeneity of the targeted region. Whole body hyperthermia is typically employed for the treatments of metastatic tumors (i.e., tumors that have spread in different parts of the human body), or some diseases like HIV. Heat is applied by different ways, either radiation or extracorporeal modality, to the whole body in the temperature range between 41 and 43°C.

Thermal ablation Thermal ablation relies on the use of high temperatures above 50°C for short duration varying in the range between few seconds up to 4–6 minutes. Such high temperatures are able to induce irreversible cell injury and ultimately tumor apoptosis and coagulative necrosis. The ablative modality is used to cure many kinds of cancer like liver, kidney, lung, bone, etc [167]. Compared to longer hyperthermia, thermo-ablative technologies appear to be more convenient. Faster tumor destruction in a minimal invasive way provides the following advantages: 1) to reduce the patients discomfort and overall costs of the treatment, 2) to shorten the hospitalization times, 3) to focus the thermal energy in the target volume preserving the healthy tissue surrounding the tumor, 4) to eliminate the effect of thermotolerance, and 5) to eliminate the influence of blood perfusion. Several energy sources have been used to overheat the tissue with different approaches, namely invasive, minimally invasive, and superficial, by means of radiofrequency (RF), ultrasound (US), microwave (MW), and light (lasers). RF create resistive heating into the tissue using electromagnetic energy delivered through needle-like electrodes directly placed into the tumor, at frequencies usually in the 46

CHAPTER 1. State of the art

range of 375–500 kHz. The passage of alternating high-frequency currents through the probe tip determines the agitation of the molecules which is converted into heat in the zone close to the electrodes tip [168]. It has received big interest as a minimally invasive technique to eradicate many types of tumor like liver, lung, bone, brain, kidney, prostate gland, pancreas, breast, etc.[152]. The use of MW for thermal cancer treatment began in the late 1970s. Current commercially available MW ablation technology utilizes electromagnetic wave frequencies of 915 MHz or 2.45 GHz to produce frictional tissue heating. Polar molecules in tissue (primarily water) are forced to continuously realign with the oscillating electric field, increasing their kinetic energy and, hence, the temperature of the tissue [169]. Unlike RF ablation, MW ablation is not an electrical current method. For this reason, it is very valuable for tissues with poor electrical conductivity like bone, kidney, and lung. Ultrasound is widely used in medicine both for medical diagnosis and therapeutic purposes such as physiotherapy, lithotripsy for kidney stone destruction, or for tissue ablation in tumor treatment in the form of high- intensity ultrasound (HIFU) [170]. With this modality, ultrasound energy can be focused precisely to ablate the target in a non-invasive or minimally invasive manner. HIFU has been widely investigated as a promising procedure for ablating benign and malignant tumors located in prostate, uterus, breast, kidney, pancreas, bone, and brain [171]. It can be employed in continuous and pulsed (p-HIFU, [172]) regimes. Laser ablation consists in the conversion of light into heat through reflection, scattering, and absorption of the incident radiation [152]. In particular, when light in the near infrared region (NIR, λ = 650 to 950 nm) is used, the therapy is known as photothermal therapy (PTT) and it is typically applied in combination of photothermal transducers (e.g., indocyanine green [173], polyaniline [174], carbon-based materials [175], etc.) that combined with NIR light are able to efficiently absorb and convert the irradiated photons into heat to raise the local temperature and thus kill cancer cells [176]. PTT represents one of the innovative therapeutic strategies which is gaining attention among cancer therapies because of its high specificity, minimal invasiveness, and precise spatial-temporal selectivity [177]. Numerous cancers can be treated by this modality such as malignant hepatic, liver, prostate, renal, pulmonary [178], brain [179], breast [180], and melanoma [181].

Physiological effects of heat Most of the physiological effects induced by temperature elevation within the tissues are related to the effects on the blood flow, which is one of the predominant factors influencing the tumor response [182]. Healthy tissues typically respond to hyperthermia with a considerable increase in blood flow as long as tissue temperature levels of 45°C, exposure times of 30–60 min, and heating rates of 0.7°C/min are not exceeded [183]. Vasodilatation occurs as a response to the increased 47

CHAPTER 1. State of the art

blood flow leading to the cooling of the temperature of the tissue by convection. Solid tumors have a more acidic and hypoxic (i.e., tissue deprivation of adequate oxygen supply) microenvironment than normal tissue which induces them to develop a new vascular supply. The latter, however, is inefficient and chaotic, resulting in preserving the factors that stimulated the neovascularization [184]. Hypoxic microenvironments are characterized by low oxygen tension, low extracellular pH, high interstitial fluid pressure, glucose deficiency, multidrug resistance, increased extracellular lactate concentration, and tendency to cause metastasis [185]. All these factors, despite contributing to the highest sensitivity of tumor cells to heat cytotoxic effect, are antagonist of the chemo and radiotherapy commonly used as a cancer treatment. Poor tumor vascularization determines a tumor blood flow which is typically inferior to that of normal tissue. The result is that vasodilation typically occurring after heat application to normal tissues is not present to the same extent in the tumor vasculature. This determines decreased convection and permits the entrapment of heat in the tumor area, leading to a temperature rise in the target zone much higher compared to the surrounding healthy tissue [186].

Cellular and molecular responses to heat Living cells use a broad complex set of metabolic pathways and regulatory balances throughout their lifespan. Effects of elevated temperature at cellular and molecular level are complex, and despite the large amount of research performed, the mechanism underlying the interaction is not fully explained. Depending on the temperature elevation, cellular responses occur on the basis of two mechanisms including 1) inactivation of protein functions and enzymatic activity, and 2) activation of signaling pathways [187]. Protein and enzymatic inactivation is responsible for heat cytotoxicity [188], [189] and radio or chemo sensitization [190] of the cells as responses to a severe heat shock (usually > 43°C), while induction of thermotolerance [191] is the dominant activating response occurring when cells are exposed to sublethal temperatures, typically ranging from 39 to 42°C. Interestingly, as suggested in [187], the activation energy for inducing heat cytotoxicity, radio/chemosensitization, and thermotolerance all fall within a similar range of 120–146 kcal/mole. The relatively high activation energy for each of these responses suggests that they all involve a heat-induced molecular transition as a trigger. Several lines of research strongly suggest that protein unfolding is the common transition that generate all three responses [187], [192]–[194]. Temperature elevation in the hyperthermic region (e.g., 40–47 °C) determines the protein unfolding and the subsequent exposition of the hydrophobic groups which may aggregate non-specifically with native proteins (Fig. 1.16, top) leading to inactivation of protein synthesis, cell cycle progression, and inhibition of DNA double strand break repair processes [187]. Other cellular 48

CHAPTER 1. State of the art

effects of mild hyperthermia include: 1) the inhibition of DNA synthesis, transcription, RNA processing, and translation, 2) increased degradation of aggregated/misfolded proteins through the proteasomal and lysosomal pathways, 3) disruption of the membrane cytoskeleton, 4) metabolic changes (e.g., uncoupling of oxidative phosphorylation) that lead to decreased levels of ATP, and 5) alterations in membrane permeability that cause increases in intracellular levels of Na+, H+, and + Ca2 . However, cellular changes induced by mild hyperthermia may be reversible through the activation of a self-defense mechanism mediated by the induction of heat shock proteins (HSP). On the contrary, temperature elevation in the ablative range (> 48 °C) is sufficient to provoke severe denaturation of proteins leading to a direct cellular death (Fig.1.16, bottom) as reported by the pioneer studies of Lepock [i.e., 169].

Figure 1.16: Thermal effects on protein folding due to hyperthermia (top), and thermal ablation (bottom).

Cytotoxicity is defined as the potential of heat to directly kill the cells. Depending on the entity of the heat shock provided at the cellular level (i.e., hyperthermia or thermal ablation) cell death may occur in different ways including apoptosis, necrosis, and catastrophic mitotic events [188], [195]. Apoptosis is defined as a “programmed cell death” which may occur normally during development and aging as well as a self-defence mechanism triggered by a wide variety of stimuli and conditions. The apoptotic machinery can become activated either via the death receptor (extrinsic) pathway or the mitochondrial (intrinsic) pathway. Both the pathways involve the activation of a group of cysteine proteases called “caspases” and a complex cascade of events that link the initiating stimuli to the final demise of the cell. Necrosis is a type of cell death occurring when cells undergo to extreme environmental conditions, adverse and excessive stimuli, or when deleterious mutations are encoded in their genetic material. Unlike apoptosis, which involves a 49

CHAPTER 1. State of the art

highly regulated and elaborated network of biochemical events and cascades, necrosis has been considered generally to be a chaotic decadent process that effects the inexorable demise of cells otherwise not destined to die [196]. Mitotic catastrophe death is a mode of cell death that results from premature or inappropriate entry of cells into mitosis and can be caused by chemical or physical stresses such as ionizing radiation, treatment with agents influencing the stability of microtubule, or various anticancer drugs [197]. Radio or chemo sensitization is defined as the enhancement in radiation or chemo-induced lethality by heat. Temperature elevation in the tumors has the ability to selectively facilitate the killing of cancerous cells that are under conditions of hypoxia, low pH, and in the S-phase of cell division, which are all conditions that render cells more resistant to radiation [198]. The mechanisms through which heat may sensitize the cells to chemo or radio therapies is typically associated to heat-induced modifications including 1) the inibition of the DNA repair, 2) the changing in tumor oxygenation, and 3) the enhancement in the membrane fluidity. Numerous studies have provided evidences of the fact that heat may induce the inhibition of the DNA double strand break repair processes, sensitising the cells to the effects of radio or chemotherapies [199], [200]. The extent of synergism between heat and radiation depends on the temperature applied, time interval between heat and radiation, and treatment sequence. It has been demonstrated that the greater effect is achieved when both modalities are applied synchronously [201]. Depending on the thermal dose heat may cause a decrease or an increase of the blood flow [202], [203]. Mild hyperthermic temperatures (< 42°C) cause an increase in the blood perfusion which has two main advantagious effects: 1) the enhancement of the vessel permeability to drugs, and 2) the inactivation of the enzymes involved in aerobic metabolism, which decreases oxygen consumption and further enhances tumor oxygenation, with consequent decrease of the population of hypoxic resistant cells. Higher temperature elevation (> 42°C), decreases blood flow allowing: 1) the achievement of higher temperature within the tumor, 2) the decrease of tumour pH which increases the sensitivity of certain drugs [204], and 3) the reduction of tumor oxygenation promoting conditions of hypoxia and nutritionally deprived environment which facilitate cells to be out of cycle. In particular, cells show enhanced sensitivity when they are in S-phase [205]. Temperature elevation at the cellular level may deeply impact the behaviour of the phospholipid bilayer and proteins composing the cellular membrane. The main effects reported regard 1) the increase of the membrane fluidity which in turn increases its permeability to drugs (i.e., chemotherapeutic Cisplatin [206] or Doxorubicin [207]), and 2) the modification of the behaviour of membrane-embedded proteins which may enhance both the cellular Cisplatin uptake and cytotoxicity [208].

CHAPTER 1. State of the art

Thermotolerance is the ability of mammalian cells, when exposed to a non-lethal shock, to acquire a transient resistance to subsequent exposures at elevated temperatures [191], [209]. It has been demonstrated that in in vitro experiments, thermotolerance may be induced by short exposure to temperatures > 43°C followed by incubation at 37°C or during continuous heating at temperatures < 43°C [210]. The phenomenon of acquired thermotolerance is transient and depends chiefly on the severity of the initial heat stress. Once the stress stimulus is removed, cells can recover their normal cellular functions. If the stress continues or is too severe, then the cells will likely die by apoptosis or necrosis. In general, up to a certain threshold beyond which the cells are killed, the greater the initial heat dose, the greater the magnitude and duration of thermotolerance. The expression of thermotolerance following heat will occur within several hours, with maximum expression generally occurring 16–18 h following the initial thermal insult and may last 3–5 days [209]. The acquisition of thermotolerance is characterized by numerous biochemical and molecular changes and it is generally associated with the existence of protein quality control response, which is one of the most conserved cytoprotective mechanisms in evolution [107]. In case of heat shock, cells overexpress chaperones and HSP that protect cellular proteins from misfolding and aggregation [107]. HSPs represent a heterogeneous family of molecular chaperones triggered by a variety of stressful stimuli, both environmental and physiological including ischemia, hypoxia, pressure overload, heavy metals, free oxygen radicals, protein kinase C, calcium increasing agents, ethanol, amino acids, inflammation, sodium arsenide, hormones, antibiotics, cytokines, infections, development, differentiation, and neurodegenerative diseases (Fig. 1.17) [107]. HSPs are classified into six main families according to their approximate molecular mass [211]: HSP100, HSP90, HSP70, HSP60, HSP40, and small HSPs when their molecular mass is < 40 kDa. The dramatic up- regulation of the HSP is a fundamental part of the heat shock response and is induced primarily by heat shock factor (HSF) [212]. The major classes of HSP induced by the heat shock response are HSP90, HSP70, HSP60, and HSP27. They have been identified as the key determinants of cell survival because they also modulate apoptosis by directly interacting with components of the apoptotic machinery [209]. In particular, HSP70 and phosphorylated HSP27 can protect cells against oxidative stress. HSP70 accumulation has been correlated to the acquisition of a thermotolerant state in mammals, amphibians, insects [213], and fishes [214]. HSP27 is a small chaperone that forms high molecular weight oligomers to trap and protect from aggregation the stress-induced misfolded proteins [215]. The ability of HSP27 to bind to client proteins depend on its level of phosphorylation which is induced by heat shock response. However, the appearance of thermotolerance during thermal treatment is undesired because it limits the efficacy of the therapy. For this reason, the right planning of the treatment is of outmost 51

CHAPTER 1. State of the art

importance to maximize heat cytotoxic effect preventing the development of heat-induced cellular resistance.

Figure 1.17: Cell stress conditions that induce the heat shock response including environmental stress, growth and development, pathophysiology, and protein conformational diseases.

1.3.2 Thermal dose Independently from the heating method used, standardized metrics are needed to quantify the relationship between thermal exposure and damage [216]. This is particularly important when comparing results from different studies, quantifying tissue-dependent sensitivities to heat, or for therapeutic purposes. The assessment of the correct thermal dosimetry is the key leading to the success of the therapy, and at the same time, it is the main issue related to clinical application of heat. Several metrics have been proposed in the years to quantify the effects of heat on biological tissues. One class of methods is based upon observed consequences due to the application of heat on tissues such as the 1) dose to produce a certain percentage decrease in liver functions [217], 2) dose to produce various serum enzyme elevations [217], 3) dose to produce changes in the hemodynamic response [217], and 4) dose to produce an arbitrary erythema score [218], [219]. Although these metrics beg the question of hyperthermia dosage, they do not provide a universal methodology that allows comparison between treatment protocols. Another class of methods is based upon measurement of a combination of different thermodynamic parameters such as: 1) total heat transferred or confined to the patient [220], 2) duration of exposure above a baseline temperature [221], 3) power administered (directly related to the temperature reached), time, and

CHAPTER 1. State of the art

intervals of treatment [222], and 4) highest or lowest temperature achieved [223]. These methods are capable of very precise quantification but are of doubtful relevance as a measure of biological response except under very restrictive conditions. A powerful thermal dose parameter should be able to quantify the dose received by the tissue during the treatment by providing a number which is related to the biological effects elicited [216]. The principal requirements of the dose include: 1) the biological response must be related to the dose in a relevant manner, 2) the dose should be a well-defined and measurable physical quantity and, 3) the dose should enable the prediction of the likely effects of a given amount of heat and the comparison between different treatments. It has been demonstrated by numerous studies (i.e., [160], [224]) that biological effects are an exponential function of temperature and time, both highly variable from one to another treatment. For this reason, numerous models based on a thermodynamic or Arrhenius-type relationship have been proposed, such as the ones by Atkinson et al. [225], Gemer et al. [226], Hahn [227], Field and Morris [228], and Sapareto and Dewey [229]. Among these models, currently, the Arrhenius damage index [228] and the cumulative equivalent minutes at 43°C (CEM43°C) [229] are used to predict the degree of thermal injury induced by ablative temperatures and to guide the treatment duration of hyperthermic therapy, respectively [230].

Kinetics of thermal damage process It has been demonstrated that thermal damage induced by heat in biological cells and/or tissues can be modeled as a first-order kinetic reaction by means of the Arrhenius equation [231], which gives the dependence of the rate constant of a chemical reaction on the absolute temperature and the activation energy:

(1.13)

This equation linearly relates the rate ( ) of temperature-dependent reaction to the frequency factor A ( ), and exponentially to the activation energy ( ), the universal gas constant R (8.314 ), and the temperature T (K). The A factor expresses the fraction of reactant molecules that possesses enough kinetic energy to react as governed by the Maxwell-

Boltzmann distribution. Depending on the magnitudes of and the temperature T, this fraction can range from 0, where no molecules have enough energy to react, to 1, where all the molecules have enough energy to react.

CHAPTER 1. State of the art

Thermal activated reactants, i.e., A and B, jump over the energy barrier to form an activated complex transition state, [AB]*, which either relaxes back to the native inactivated single reactants or progresses to form products molecules (Fig. 1.18). The sequence of formation is:

k b

k (1.14) a

Figure 1.18: Energy-state diagram. Reactants surmount energy barrier to transition to products [232].

This model applies well to the study of protein denaturation which represents the trigger of all the cellular responses to heat [188]. Thermal damage in tissue is considered as a unimolecular process: cellular transition from an initial state (native state) at a certain concentration of molecules C to the damaged (denatured) state occurs following the temperature elevation. The initial concentration C, surmounts an energy barrier, ΔH*, to become activated molecules, C*, at a rate , some of which relax back to inactivated native state molecules at rate , and others become -1 denatured molecule at rate (Fig. 1.19). ΔH*is defined as activation enthalpy (J∙mol ) and it can be approximated to the activation energy for first order reactions [232].

CHAPTER 1. State of the art

Figure1.19: Energy-state diagram for unimolecular reaction with energy barrier [232].

The kinetic parameters A and Ea are highly variable depending on the heat sensitivity of specific tissue or cell line, the assay used to characterize the cell survival (i.e., clonogenic assays, cell death assay with tripan blue, propidium dye, etc.) or the tissue damage (i.e., assays with different end points such as tissue necrosis), the exposure conditions, and the time when assessment of injury is performed. Typical values of A and are reviewed in [232]–[234]. The experimental procedure that may be used to obtain the kinetic parameters for in in vitro cells is briefly defined hereafter. First, cell survival curves have to be determined by exposing in vitro cells to a defined temperature range and measuring their viability, after exposure, at specified time intervals. This allows the definition of dose-effect curves by plotting the rate of cell death against the duration of the heat exposure. In order to avoid complex non-linear data fitting, isothermal heating is preferred to non- isothermal one [235]. Second, the cell injury rate , has to be determined. At a given temperature, cell survival S is defined as:

(1.15) where S (%) is the cell survival after heating at temperature T (K) in a certain time interval t (s), and

( ) is the injury rate as defined by equation (1.13). Therefore, can be determined by fitting equation (1.15) to experimental data of cell survival versus isothermal heating time at various constant temperatures (Fig. 1.20a). Third, and A can be retrieved from the values of (obtained in the previous step) versus temperature. In particular, the relationship between the natural logarithm of versus the inverse of temperature (in Kelvin) can be plotted and fit using the linear equation derived from equation (1.13), but taking the natural logarithm on both sides of the equation:

(1.16)

Ea and A can then be determined from the slope and intercept of the linear fit (Fig. 1.20b).

CHAPTER 1. State of the art

Figure 1.20: (a) Determination of cell injury rate (k in the figure) at various temperatures by fitting equation (1.15) to experimental cell survival data. (b) Determination of kinetics parameters by linear fitting using equation (1.13) from the slope and the intercept of the line [235].

Thermal damage model The Arrhenius formulation of thermal damage process was first employed in 1947 by Henriquez and Moritz in a number of works [236]–[239] aimed to quantify the extent of thermal damage of skin burns in pigs. In their works burns were created in pig skin by flowing water over the skin surface at constant temperature, and the relative damage was calculated in the form of a dimensionless parameter Ω(τ):

(1.17)

where τ (s) is the duration of the exposure, A is the frequency factor, and , is the reaction rate. Equation (1.17) can be recast by considering that thermal damage is a first-order unimolecular reaction where the rate of disappearance of the reactant is described by a Bernoulli differential equation:

(1.18)

with C is the concentration of reactant. The solution of the differential equation is determined by:

(1.19)

CHAPTER 1. State of the art

where C(0) represents the number of undamaged cells before the application of heat and is the number of the undamaged cells at time τ (s). Therefore, the physical significance of the damage index, Ω(τ), is the logarithm of the ratio of the original concentration of native cells to the remaining ones at time τ (s):

(1.20)

At any time, the thermal damage probability can be predicted by:

P(%) = 1 – C(τ)=100 (1 – ) (1.21)

The original studies [236]–[239] quantified to identify various injury thresholds of skin. was assigned to a “first degree burning” or the minimum conditions to obtain irreversible epidermal injury, to a “second degree burning” or a complete transepidermal necrosis, and Ω = to a “third degree burning” or full-thickness burn which completely involves dermis. It is possible to correlate these thresholds to the prediction of the effective thermal damage by using equation (1.21). For example, Ω values of 0.53, 1, and 104 correspond, respectively to a cellular damage 41.1%, 63.3%, and 100% complete (i.e., 58.9%, 36.7%, and 0% survival, respectively). However, these thresholds may differ from one to another tissue and/or cell line. Therefore, considering the thermal damage in terms of remaining undamaged tissue constituent, C(τ) (i.e., by using left part of equation (1.20)) appears more appropriate. The advantages are: 1) the model directly reveals what is happening in tissues or cells, and 2) the prediction of C(τ) can be directly compared between numerical model results and histological analyses or functional imaging based on a quantitative measure [232].

CEM43°C model The CEM43°C model was proposed in 1984 by Sapareto and Dewey [240] to calculate the thermal isoeffect dose achieved during hyperthermic treatments. It enables to convert any time- temperature exposure to an equivalent number of minutes of heating at a reference temperature, commonly 43 °C, using the following formula:

(1.22)

CHAPTER 1. State of the art

where t (min) is the duration of the exposure, T (°C) is the actual temperature of the target tissue,

is the factor to compensate for a one degree temperature change above or below 43°C; in the discretized form, n represents the number of intervals in which the duration of the exposure has been divided, (°C) the average temperature in the interval (min), and 43°C the reference temperature. The model is based on dose-response cell survival curves. It has been demonstrated that the rate of cell killing is exponential and it is function of the temperature and duration of the exposure [160], [241]. The slope of the curves is determined by the cellular sensitivity to heat (i.e., the strongest heat resistance of the human cells results in the shallow slope of their dose-response curves compared to the ones obtained in rodents, at any defined temperature [242]). At low temperatures, i.e., < 43°C and short exposure times, the survival curves present a shoulder (indicative of the development of thermotolerance), which changes into a constant rate of cell death for temperatures typically > 43 °C (Fig. 1.21). The temperature of 43°C typically corresponds to a breakpoint in the cell survival curves and it likely represents the upper limit at which thermal tolerance can be induced in mammalian cells [193]. This explains why it has been chosen as the reference temperature for the CEM43°C model. However, numerous studies showed that the breakpoint may vary from 43 to 47°C [241] depending on the cell line considered (i.e., stronger resistance of the HTB-66 cells correspond to a breakpoint of 43.5°C while it appears at 43°C in CHO-10B cells [Fig. 1.21, [242]] ). The different cellular behaviour above and below the breakpoint is taken into account in the model by the parameter which is experimentally derived from the Arrhenius law. It depends much more strongly on the activation energy than it does on the temperature, hence it is typically considered constant during the exposure [232]. In the original formulation of the model [240], = 0.5 for T > 43°C, while = 0.25 for T < 43°C, indicating that the rate of cell death doubles for every 1°C rise above the breakpoint, and decreases by a factor of 4 per degree temperature decrease below the breakpoint. These values have been derived from the Chinese Hamster Ovary (CHO) cell experiments [242]. It is fundamental to correctly characterize the value depending on the specific application due to its large variability as a function of the type of study (i.e., in vivo or in vitro), the species (i.e., human or rodent), or the cell or tissue of interest [241].

CHAPTER 1. State of the art

(a) (b) Figure 1.21: (a) Survival of Chinese hamster ovary cells (CHO) heated over a range of temperatures. (b) Survival of a human melanoma (HTB-66) cell heated to similar ranges of temperatures [242].

The CEM43°C has been largely employed in the clinical application to define the correct duration of the treatment able to achieve the cancer destruction. Numerous data have been collected to establish the threshold for thermal damage of tissue as extensively reviewed in [241], [243]. However, numerous limitations are related to the model as widely explained by Pearce [232]. The first limitation is associated to the difficult issue to obtain precise temperature measurements during the treatment which directly impacts the accuracy of the prediction of the thermal dose. Second, the Arrhenius model is not functionally capable of representing the shoulder region (e.g., the cellular thermotolerance) of experimental survival curves for many cellular processes. Third, the model describes only a single constant irreversible reaction of the total thermal damage process leading to loss of clonogenicity as a surrogate for cell death. Nevertheless, the process leading to cell death is a complex cascade of multiple pathways involving thermal effects on proteins, cell membranes, nucleus, cytoskeleton, and activation (or inactivation) of cellular functions. Fourth, it does not take into account physiological changes induced in tissue by heat, i.e., the effect of blood perfusion. Fifth, the model does not include the effect of sensitization by enhanced oxygenation or by physical or chemical agents like drugs. Sixth, it does not include step-down heating or different thermal sensitivity of cells to pH changes [244]. Finally, it predicts cell death at low temperatures for very long exposure durations, although these temperature are known to be not thermally hazardous. Despite all these numerous limitations and the many attempts to provide complex alternatives (i.e., introducing multi-parameter and stochastic models [232]), the CEM43°C appears to be the

CHAPTER 1. State of the art

optimal practical solution to relate the temperature-time history in an exponential manner to the biological response during hyperthermia.

1.3.3 Importance of pulsed waveform in heat-induced cellular responses Conventional thermo-treatments typically provide continuous induced heating of the tissue by means of ultrasounds or electromagnetic waves in a large band of frequencies. Despite the large amount of research performed in the field, which has permitted the great advance of the therapies, numerous drawbacks are associated to the use of traditional hyperthermic treatments such as 1) under-dosed targets, 2) damage of the healthy tissue, and 3) long treatment times. Numerous studies have been performed on minimally invasive treatment of tumors by means of short electric pulses (Fig. 1.22), i.e., the so called electroporation [245]. The application of short electrical pulses of the order of µs [246]or ns [247], [248] can induce transmembrane voltage across the cell plasma membrane, resulting in its permeabilization to agents that otherwise could not transfer into it [249]. Numerous electroporation-based medical technologies have been developed in particular for the cancer treatment including 1) electrochemotherapy (ECT), 2) nanosecond pulsed electric fields (nsPEFs), and 3) irreversible electroporation (IRE). ECT is a safe, efficient, local, and non-thermal tumor ablation modality for the treatment of solid tumors that uses cell-permeabilizing electrical pulses to enhance the activity of a non-permeant (bleomycin) or low permeant (cisplatin) anticancer drug with a very high intrinsic cytotoxicity [245], [250]. Numerous studies have demonstrated that high powered, nsPEF may induce cell death by apoptosis and necrosis in vitro and reduce the size of tumours both in animal models and in humans [245], [251]–[253]. IRE is a new ablation procedure that provokes the permanent permeabilization of the cells resulting in their death without causing any detrimental thermal effect and without the need of adjuvant drugs [254].

CHAPTER 1. State of the art

Figure1.22: Typical parameters of high-voltage electric pulse in the nanosecond duration range: pulse repetition rate, peak voltage, and energy [248].

However, despite the success of these therapies and the large amount of research already performed in the electroporation direction, new tendencies in the oncological field seem to be oriented to the use of pulsed heating as an advantageous tool for the optimization of the treatment efficacy. As shown in Figure 1.23 high peak temperature necessary for the killing of the cancerous cells is rapidly achieved in the warm-up phase of the pulse. After exposure and after reaching the threshold temperature (T0, in Fig. 1.23) temperature decays due to losses through thermal diffusion because of the colder neighbouring tissues. In spite of the many advantages possibly deriving from the use of short thermal pulses, these modalities have been less explored. A literature review of the main studies performed in the field of the pulsed thermal therapies is presented hereafter. Following sub-sections are divided as a function of the likely advantages introduced by pulsed thermo-treatments. The review takes into account studies exploiting a wide frequency band between 300 kHz and 400 THz.

CHAPTER 1. State of the art

Figure 1.23: Example of thermal pulses with different temperature rises; 505 and 515 refers to the lowest and fastest temperature rise, respectively [255].

Highly localized heating The goal of hyperthermia as well as thermal ablation is to destroy the cancerous tissue without damaging the healthy one. However, when continuous heating is delivered for long time (i.e., during hyperthermia) the overheating of the tissue surrounding the tumor is unavoidable. On the contrary, if heat is generated as form of short thermal pulses, the phenomenon of heat diffusion is limited, and therefore, heat remains localized only in the zone of interest [256]–[258]. Figure 1.24 presents an example of the rapid decrease of the surface temperature with the distance after the application of 500 pulses of 200 ns and repetition rate of 1 kHz [258]. Duration of the thermal pulse strongly influences the thermal distribution in the healthy tissue; if the duration of the pulse is shorter than the thermal diffusion time, the distribution of the thermal energy is confined within the irradiated zone, while pulses longer than the thermal relaxation time lead to diffusion of heat outside the affected zone causing the possible heating of the neighboring healthy tissue [256]. Recent technologies allow the generation of short thermal pulses throughout the electromagnetic spectrum. Among the others, ultra short (picosecond and femtosecond) laser pulses have been recognized as powerful tools to provide high precision and minimal collateral thermal tissue damage for the selective destruction of cancer [259] as extensively reviewed in [260], and more recently in [261].

CHAPTER 1. State of the art

Figure 1.24: Variation of surface temperature with change in power of laser beam (focused at tissue surface) [258].

Homogeneity of the thermal dose and independence from blood perfusion Large inhomogeneity of the thermal properties of the tissues results in an equal large inhomogeneity of their heating. This means that during the treatment, the tumor does not receive the same amount of heat resulting in under-dosed targets possibly leading to the failure of the therapy. Moreover, the cooling effect of the vascular perfusion has a great impact on the dose distribution within the tissue, generating large thermal gradients and cold spots within the zone of interest. Therefore the development of perfusion-independent treatments has attracted the attention of many researchers [262], [263]. For example, Billard et al. [263] experimentally demonstrated that the effects of blood perfusion are negligible for short pulses ≤ 2 s (Fig. 1.25).

Figure 1.25: Peak temperature vs. pulse length for total flows of 0, 5, 43, and 130 ml/min into in vitro kidney [263]. 63

CHAPTER 1. State of the art

Moreover, numerous studies have been performed to characterize the feasibility of thermal pulses-based therapies [264] by analysing the dependence of the temperature homogeneity on the i) heating pulse duration [265], ii) focal spacing between the pulses [266], iii) size of the focal zone [262], iv) desired thermal dose [267], and v) energy and power deposition within the irradiated tissue [268], [269].

Larger extension of the thermal damage Pulsed heating can significantly reduce the heat build-up in tissues through non-continuous energy deposition with lower temporal average intensity. When the duration of the pulse is sufficiently short (typically shorter than the thermal diffusion time), the temperature elevation within the surrounding healthy tissue in the off period of the pulse is negligible. This means that high power can be applied during the on period of the pulse, enabling deeper heat penetration and greater tissue lesions [267]–[270] than the equivalent lesions obtained through the delivery of lower continuous power for the same energy (Fig. 1.26).

Figure 1.26: Cross-sectional area of the ablation zones in vivo liver for CW and PW MW exposure with the same average power (25 W). Different stages of tissue desiccation were observed within the ablation zone: little desiccation (orange arrow), more complete desiccation or tissue necrosis (black arrow), and charring (white arrow) [269].

Energy delivery as form of pulses provided more rapid heating and achieved higher temperatures and thermal doses than continuous heating for equivalent continuous average power [269], [271] (Fig. 1.27). 64

CHAPTER 1. State of the art

Figure 1.27: Thermal dose comparison of CW laser and pulsed laser for 100 pulses with average power of 1.3 W [258].

A direct correlation has been found between the extent of the thermal injury and 1) the pulse duration and the related peak power (e.g., shorter pulses associated to higher power [Fig. 1.28], are more efficient in terms of ablative action compared to longer but less powerful pulses for the same energy delivered [272]), 2) the duty cycle, defined as the ratio of the pulse duration to the total pulse period (e.g., shorter duty cycles appear to provide greater benefit [172], [269]) and, 3) temperature rise [273].

Figure 1.28: Peak power delivered as a function of the pulse duration (s) [268].

CHAPTER 1. State of the art

The improved efficacy of the treatments results in the shortening of the therapy duration reducing the treatment times and the related expenses, the normal tissue pain and damage which is beneficial for the patient comfort, and the development of the undesired cellular thermotolerance.

Bibliography

[1] M. A. Rojavin and M. C. Ziskin, “Medical application of millimetre waves,” QJM Mon. J. Assoc. Physicians, vol. 91, no. 1, pp. 57–66, Jan. 1998. [2] W. Snyder, M. Cook, E. Nasset, L. Karhausen, G. Howells, and Tipton IH, “Report of the task group on reference man,” Int. Comm. Radiol. Prot., no. 23, pp. 46–49, 1974. [3] A. Baroni, E. Buommino, V. De Gregorio, E. Ruocco, V. Ruocco, and R. Wolf, “Structure and function of the epidermis related to barrier properties,” Clin. Dermatol., vol. 30, no. 3, pp. 257– 262, May 2012. [4] M. Geerligs, Skin layer mechanics. Technische Universiteit Eindhoven, 2010. [5] J. Sandby-Møller, T. T. Poulsen, and H. C. O. Wulf, “Epidermal thickness at different body sites: relationship to age, gender, pigmentation, blood content, skin type and smoking habits.,” Acta Derm. Venereol., vol. 83, no. 6, pp. 410–413, 2003. [6] National Cancer Institute, “SEER training modules, layers of the skin.” [Online]. Available: https://training.seer.cancer.gov/melanoma/anatomy/layers.html. [7] M. C. Ziskin, S. I. Alekseev, K. R. Foster, and Q. Balzano, “Tissue models for RF exposure evaluation at frequencies above 6 GHz,” Bioelectromagnetics, vol. 39, no. 3, pp. 173–189, Apr. 2018. [8] M. Gaur, M. Dobke, and V. Lunyak, “Mesenchymal stem cells from adipose tissue in clinical applications for dermatological indications and skin aging,” Int J Mol Sci, 2017. [9] Z. Apalla, A. Lallas, E. Sotiriou, E. Lazaridou, and D. Ioannides, “Epidemiological trends in skin cancer,” Dermatol. Pract. Concept., vol. 7, no. 2, pp. 1–6, Apr. 2017. [10] American Cancer Society, “Information and resources about for cancer: breast, colon, lung, prostate, skin.” [Online]. Available: https://www.cancer.org/. [11] U. Leiter, T. Eigentler, and C. Garbe, “Epidemiology of skin cancer,” Adv. Exp. Med. Biol., vol. 810, pp. 120–140, 2014. [12] American Cancer Society, “Key statistics for melanoma skin cancer.” [Online]. Available: https://www.cancer.org/content/cancer/en/cancer/melanoma-skin-cancer/about/key-statistics/. [13] Melanoma Research Foundation (MRF), “What is Melanoma?” [Online]. Available: https://www.melanoma.org/understand-melanoma/what-is-melanoma. [14] D. M. Gress et al., “Principles of cancer staging,” in AJCC Cancer Staging Manual, Cham: Springer International Publishing, 2017, pp. 3–30. [15] P. V. Dickson and J. E. Gershenwald, “Staging and prognosis of cutaneous melanoma,” Surg. Oncol. Clin. N. Am., vol. 20, no. 1, pp. 1–17, Jan. 2011. [16] J. E. Gershenwald, “8th Edition AJCC Melanoma Staging System,” p. 29, 2018. [17] S. Moqadam, P. Grewal, Z. Haeri, P. Ingledew, K. Kohli, and F. Golnaraghi, “Cancer detection based on electrical impedance spectroscopy: a clinical study,” J. Electr. Bioimpedance, vol. 9, no. 1, pp. 17–23, Aug. 2018. [18] S. I. Alekseev and M. C. Ziskin, “Biological effects of millimeter and submillimeter waves,” in Handbook of biological effects of electromagnetic fields. Biological and medical aspects of electromagnetic field, Fourth Edn., Boca Raton, FL: CRC Press, 2018, pp. 179–241. [19] K. R. Foster and H. P. Schwan, “Dielectric properties of tissues,” in Handbook of biological effects of electromagnetic fields. Biological and medical aspects of electromagnetic field, Second Edn., Boca Raton, FL: CRC Press, 1996, pp. 25-. [20] J. K. Seo, T. K. Bera, H. Kwon, and R. Sadleir, “Effective admittivity of biological tissues as a coefficient of elliptic PDE,” Comput. Math. Methods Med., vol. 2013, pp. 1–10, 2013. [21] K. R. Foster and H. P. Schwan, “Dielectric properties of tissues and biological materials: a critical review,” Crit. Rev. Biomed. Eng., vol. 17, no. 1, pp. 25–104, 1989. [22] H. P. Schwan, “Field interaction with biological matter,” Ann. N. Y. Acad. Sci., vol. 303, pp. 198–216, Dec. 1977.

Bibliography

[23] S. Gabriel, R. W. Lau, and C. Gabriel, “The dielectric properties of biological tissues: III. Parametric models for the dielectric spectrum of tissues,” Phys. Med. Biol., vol. 41, no. 11, pp. 2271–2293, Nov. 1996. [24] M. C. Ziskin, “Physiological mechanisms underlying millimeter wave therapy,” in Biolectromagnetics Current Concepts, 2006, pp. 241–251. [25] O. P. Gandhi and A. Riazi, “Absorption of millimeter waves by human beings and its biological implications,” IEEE Trans. Microw. Theory Tech., vol. 34, no. 2, pp. 228–235, Feb. 1986. [26] C. M. Alabaster, “Permittivity of human skin in millimetre wave band,” Electron. Lett., vol. 39, no. 21, pp. 1521–1522, Oct. 2003. [27] S. Gabriel, R. W. Lau, and C. Gabriel, “The dielectric properties of biological tissues: III. Parametric models for the dielectric spectrum of tissues,” Phys. Med. Biol., vol. 41, no. 11, p. 2271, 1996. [28] S. I. Alekseev and M. C. Ziskin, “Human skin permittivity determined by millimeter wave reflection measurements,” Bioelectromagnetics, vol. 28, no. 5, pp. 331–339, Jul. 2007. [29] N. Chahat, M. Zhadobov, and R. Sauleau, “Broadband tissue-equivalent phantom for BAN applications at millimeter waves,” IEEE Trans. Microw. Theory Tech., vol. 60, no. 7, pp. 2259– 2266, Jul. 2012. [30] N. Chahat, M. Zhadobov, R. Sauleau, and S. I. Alekseev, “New method for determining dielectric properties of skin and phantoms at millimeter waves based on heating kinetics,” IEEE Trans. Microw. Theory Tech., vol. 60, no. 3, pp. 827–832, Mar. 2012. [31] H. Hwang, J. Yim, J.-W. Cho, C. Cheon, and Y. Kwon, “110 GHz broadband measurement of permittivity on human epidermis using 1 mm coaxial probe,” in IEEE MTT-S International Microwave Symposium Digest, 2003, Philadelphia, PA, USA, 2003, vol. 1, pp. 399–402. [32] L. F. Chen, C. K. Ong, C. P. Neo, V. V. Varadan, and V. K. Varadan, Microwave electronics: measurements and materials characterization. John Wiley & Sons, 2004. [33] V. Suntzeff and C. Carruthers, “The Water content in the epidermis of mice undergoing carcinogenesis by methylcholanthrene,” Cancer Res., vol. 6, no. 10, pp. 574–577, Oct. 1946. [34] P. Mehta, K. Chand, D. Narayanswamy, D. G. Beetner, R. Zoughi, and W. V. Stoecker, “Microwave reflectometry as a novel diagnostic tool for detection of skin cancers,” IEEE Trans. Instrum. Meas., vol. 55, no. 4, pp. 1309–1316, Aug. 2006. [35] M. Gniadecka et al., “Melanoma diagnosis by raman spectroscopy and neural networks: structure alterations in proteins and lipids in intact cancer tissue,” J. Invest. Dermatol., vol. 122, no. 2, pp. 443–449, Feb. 2004. [36] R. Aminzadeh, A. A. Shishegar, and M. Saviz, “Numerical and experimental assessment of millimeter-wave reflectometry for non-invasive diagnosis of skin cancers,” 8th European Conference on Antenna and Propagation (EuCAP 2014), The Hague, The Netherlands, 06-Apr- 2014. [37] R. Aminzadeh, M. Saviz, and A. A. Shishegar, “Dielectric properties estimation of normal and malignant skin tissues at millimeter-wave frequencies using effective medium theory,” in 2014 22nd Iranian Conference on Electrical Engineering (ICEE), 2014, pp. 1657–1661. [38] M. Zhadobov, C. Leduc, A. Guraliuc, N. Chahat, and R. Sauleau, “Antenna/human body interactions in the 60 GHz band: state of knowledge and recent advances,” in Advances in body- centric wireless communication: applications and state-of-the-art, Institution of Engineering and Technology (IET), Digital library, 2016, pp. 97–142. [39] M. Zhadobov, N. Chahat, R. Sauleau, C. Le Quement, and Y. Le Drean, “Millimeter-wave interactions with the human body: state of knowledge and recent advances,” Int. J. Microw. Wirel. Technol., vol. 3, no. 02, pp. 237–247, Apr. 2011. [40] T. L. Bergman, A. S. Lavine, F. P. Incropera, and D. P. Dewitt, Fundamentals of heat and mass transfer, Seventh Edn. Hoboken, NJ: John Wiley, 2007.

Bibliography

[41] F. Xu, T. J. Lu, K. A. Seffen, and E. Y. K. Ng, “Mathematical modeling of skin bioheat transfer,” Appl. Mech. Rev., vol. 62, no. 5, pp. 050801-050801–35, Jul. 2009. [42] A. Guraliuc, M. Zhadobov, and R. Sauleau, “Dosimetric aspects related to the human body exposure to millimeter waves,” MiWaveS 2020 Heterog. Wirel. Netw. Millim.-Wave Small-Cell Access Backhauling, 2014. [43] W. J. Ellison, “Permittivity of pure water, at standard atmospheric pressure, over the frequency range 0–25 THz and the temperature range 0–100°C,” J. Phys. Chem. Ref. Data, vol. 36, no. 1, pp. 1–18, Feb. 2007. [44] S. I. Alekseev, A. A. Radzievsky, M. K. Logani, and M. C. Ziskin, “Millimeter wave dosimetry of human skin,” Bioelectromagnetics, vol. 29, no. 1, pp. 65–70, Jan. 2008. [45] Y. Feldman et al., “The electromagnetic response of human skin in the millimetre and submillimetre wave range,” Phys. Med. Biol., vol. 54, no. 11, pp. 3341–3363, Jun. 2009. [46] H. H. Pennes, “Analysis of tissue and arterial blood temperatures in the resting human forearm,” J Appl Physiolo, vol. 1, no. 2, 1948. [47] K. R. Foster, H. N. Kritikos, and H. P. Schwan, “Effect of surface cooling and blood flow on the microwave heating of tissue,” IEEE Trans. Biomed. Eng., vol. 25, no. 3, pp. 313–316, May 1978. [48] J. D. Hardy, “Temperature regulation, exposure to heat and cold, and effects of hyperthermia,” in Therapeutic Heat and Cold, Third Edn., Lehmann JF Baltimore/London: Williams and Wilkins, 1982, pp. 172–198. [49] S. I. Alekseev, A. A. Radzievsky, I. Szabo, and M. C. Ziskin, “Local heating of human skin by millimeter waves: effect of blood flow,” Bioelectromagnetics, vol. 26, no. 6, pp. 489–501, Sep. 2005. [50] S. I. Alekseev and M. C. Ziskin, “Influence of blood flow and millimeter wave exposure on skin temperature in different thermal models,” Bioelectromagnetics, vol. 30, no. 1, pp. 52–58, 2009. [51] S. I. Alekseev and M. C. Ziskin, “Millimeter-wave absorption by cutaneous blood vessels: a computational study,” IEEE Trans. Biomed. Eng., vol. 56, no. 10, pp. 2380–2388, Oct. 2009. [52] A. Kanezaki, S. Watanabe, A. Hirata, and H. Shirai, “Theoretical analysis for temperature elevation of human body due to millimeter wave exposure,” in 2008 Cairo International Biomedical Engineering Conference, 2008, pp. 1–4. [53] A. Kanezaki, A. Hirata, S. Watanabe, and H. Shirai, “Parameter variation effects on temperature elevation in a steady-state, one-dimensional thermal model for millimeter wave exposure of one- and three-layer human tissue,” Phys. Med. Biol., vol. 55, no. 16, pp. 4647– 4659, Aug. 2010. [54] L. Zilberti, A. Arduino, O. Bottauscio, and M. Chiampi, “Parametric analysis of transient skin heating induced by terahertz radiation,” Bioelectromagnetics, vol. 35, no. 5, pp. 314–323, Jul. 2014. [55] L. Zilberti, A. Arduino, O. Bottauscio, and M. Chiampi, “A model to analyze the skin heating produced by millimeter and submillimeter electromagnetic waves,” in 2013 International Conference on Electromagnetics in Advanced Applications (ICEAA), 2013, pp. 895–898. [56] K. Sasaki, M. Mizuno, K. Wake, and S. Watanabe, “Monte Carlo simulations of skin exposure to electromagnetic field from 10 GHz to 1 THz,” Phys. Med. Biol., vol. 62, no. 17, pp. 6993–7010, Aug. 2017. [57] T. Samaras and N. Kuster, “Theoretical evaluation of the power transmitted to the body as a function of angle of incidence and polarization at frequencies >6 GHz and its relevance for standardization,” Bioelectromagnetics, vol. 40, no. 2, pp. 136–139, Feb. 2019. [58] “Non-Fourier based thermal-mechanical tissue damage prediction for thermal ablation: Bioengineered: Vol 8, No 1.” [Online]. Available:

Bibliography

https://www.tandfonline.com/doi/full/10.1080/21655979.2016.1227609. [Accessed: 26-Aug- 2019]. [59] IEEE Std 1528-2003, “IEEE Recommended practice for determining the peak spatial- average Specific Absorption Rate (SAR) in the human head from wireless communications devices: measurement techniques,” IEEE Std 1528-2003, pp. 1–120, Dec. 2003. [60] “MCL-T broadband tissue equivalent liquid: 30 MHz to 6 GHz,” MCL-T, London, U.K. [Online]. Available: http://www.mcluk.org/ pdfs/bbl.pdf.” [61] A. W. Guy, “Analyses of electromagnetic fields induced in biological tissues by thermographic studies on equivalent phantom models,” IEEE Trans. Microw. Theory Tech., vol. 19, no. 2, pp. 205–214, Feb. 1971. [62] K. Ito, K. Furuya, Y. Okano, and L. Hamada, “Development and characteristics of a biological tissue-equivalent phantom for microwaves,” Electron. Commun. Jpn. Part Commun., vol. 84, no. 4, pp. 67–77, Apr. 2001. [63] T. Takimoto, T. Onishi, K. Saito, M. Takahashi, S. Uebayashi, and K. Ito, “Characteristics of biological tissue equivalent phantoms applied to UWB communications,” Electron. Commun. Jpn. Part Commun., vol. 90, no. 5, pp. 48–55, 2007. [64] J. Yonebayashi, S. Takamatsu, K. Saito, M. Takahashi, and K. Ito, “Development of dynamic phantom for evaluation of breath detection Doppler radar,” presented at the 32nd Annu. Bioelectromagn. Soc. Meeting, Seoul, Korea, 2010, pp. 297–299. [65] A. A. Shishegar, R. Aminzadeh, and M. Saviz, “Theoretical and experimental broadband tissue-equivalent phantoms at microwave and millimetre-wave frequencies,” Electron. Lett., vol. 50, no. 8, pp. 618–620, Apr. 2014. [66] S. Castello-Palacios et al., “Elaboration of simple gel phantoms for 5G/mm wave communications,” in 2018 IEEE 29th Annual International Symposium on Personal, Indoor and Mobile Radio Communications (PIMRC), Bologna, Italy, 2018, pp. 1215–1219. [67] J. Lacik, V. Hebelka, J. Velim, Z. Raida, and J. Puskely, “Wideband skin-equivalent phantom for V- and W-band,” IEEE Antennas Wirel. Propag. Lett., vol. 15, pp. 211–213, 2016. [68] S. Di Meo et al., “Tissue-mimicking materials for breast phantoms up to 50 GHz,” Phys. Med. Biol., Jan. 2019. [69] N. Chahat, M. Zhadobov, S. Alekseev, and R. Sauleau, “Human skin-equivalent phantom for on-body antenna measurements in 60 GHz band,” Electron. Lett., vol. 48, no. 2, pp. 67–68, Jan. 2012. [70] C. Leduc and M. Zhadobov, “Thermal model of electromagnetic skin-equivalent phantom at millimeter waves,” IEEE Trans. Microw. Theory Tech., vol. 65, no. 3, pp. 1036–1045, Mar. 2017. [71] T. Kobayashi, T. Nojima, K. Yamada, and S. Uebayashi, “Dry phantom composed of ceramics and its application to SAR estimation,” IEEE Trans. Microw. Theory Tech., vol. 41, no. 1, pp. 136–140, 1993. [72] H. Tamura, Y. Ishikawa, T. Kobayashi, and T. Nojima, “A dry phantom material composed of ceramic and graphite powder,” IEEE Trans. Electromagn. Compat., vol. 39, no. 2, pp. 132– 137, May 1997. [73] Y. Nikawa, M. Chino, and K. Kikuchi, “Soft and dry phantom modeling material using silicone rubber with carbon fiber,” IEEE Microw. Therory Tech. Soc., vol. 44, no. 10, pp. 1949– 1953, 1996. [74] J. T. Chang, M. W. Fanning, P. M. Meaney, and K. D. Paulsen, “A conductive plastic for simulating biological tissue at microwave frequencies,” IEEE Trans. Electromagn. Compat., vol. 42, no. 1, pp. 76–81, Feb. 2000. [75] A. R. Guraliuc, M. Zhadobov, O. D. Sagazan, and R. Sauleau, “Solid phantom for body- centric propagation measurements at 60 GHz,” IEEE Trans. Microw. Theory Tech., vol. 62, no. 6, pp. 1373–1380, Jun. 2014.

Bibliography

[76] SPEAG, Schmid & Partner Engineering AG, “POPEYE.” [Online]. Available: https://speag.swiss/news-events/news/measurement/new-product-popeye/. [77] A. Paffi et al., “Considerations for developing an RF exposure system: a review for in vitro biological experiments,” IEEE Trans. Microw. Theory Tech., vol. 58, no. 10, pp. 2702–2714, Oct. 2010. [78] I. Szabo, M. A. Rojavin, T. J. Rogers, and M. C. Ziskin, “Reactions of keratinocytes to in vitro millimeter wave exposure,” Bioelectromagnetics, vol. 22, no. 5, pp. 358–364, Jul. 2001. [79] I. Szabo, M. R. Manning, A. A. Radzievsky, M. A. Wetzel, T. J. Rogers, and M. C. Ziskin, “Low power millimeter wave irradiation exerts no harmful effect on human keratinocytes in vitro,” Bioelectromagnetics, vol. 24, no. 3, pp. 165–173, Apr. 2003. [80] P. H. Siegel and V. Pikov, “Impact of low intensity millimetre waves on cell functions,” Electron. Lett., vol. 46, no. 26, pp. s70–s72, Dec. 2010. [81] M. Zhadobov et al., “Evaluation of the potential biological effects of the 60-GHz millimeter waves upon human cells,” IEEE Trans. Antennas Propag., vol. 57, no. 10, pp. 2949–2956, Oct. 2009. [82] C. L. Quément et al., “Whole-genome expression analysis in primary human keratinocyte cell cultures exposed to 60 GHz radiation,” Bioelectromagnetics, vol. 33, no. 2, pp. 147–158, 2012. [83] A. J. Haas, Y. Le Page, M. Zhadobov, A. Boriskin, R. Sauleau, and Y. Le Dréan, “Impact of 60-GHz millimeter waves on stress and pain-related protein expression in differentiating neuron-like cells: TRPV Expression Under MMW Exposure,” Bioelectromagnetics, vol. 37, no. 7, pp. 444–454, Oct. 2016. [84] M. Zhadobov, R. Sauleau, V. Vie, M. Himdi, L. L. Coq, and D. Thouroude, “Interactions between 60-GHz millimeter waves and artificial biological membranes: dependence on radiation parameters,” IEEE Trans. Microw. Theory Tech., vol. 54, no. 6, pp. 2534–2542, Jun. 2006. [85] R. K. Sypniewska et al., “Protein changes in macrophages induced by plasma from rats exposed to 35 GHz millimeter waves,” Bioelectromagnetics, vol. 31, no. 8, pp. 656–663, 2010. [86] M. Zhadobov, R. Sauleau, R. Augustine, C. Le Quément, Y. Le Dréan, and D. Thouroude, “Near-field dosimetry for in vitro exposure of human cells at 60 GHz,” Bioelectromagnetics, vol. 33, no. 1, pp. 55–64, Jan. 2012. [87] M. Zhadobov, R. Sauleau, D. Thouroude, C. N. Nicolaz, C. L. Quément, and Y. L. Dréan, “Near-field electromagnetic dosimetry for in vitro studies at millimeter waves,” in Proceedings of the Fourth European Conference on Antennas and Propagation, 2010, pp. 1–3. [88] Q. Chen, Q. L. Zeng, D. Q. Lu, and H. Chiang, “Millimeter wave exposure reverses TPA suppression of gap junction intercellular communication in HaCaT human keratinocytes,” Bioelectromagnetics, vol. 25, no. 1, pp. 1–4, 2004. [89] J. X. Zhao, D. G. Liu, H. K. Zhang, and D. Q. Lu, “Influence of culture dish configuration upon MMW irradiation distribution on cell monolayer,” Int. J. Infrared Millim. Waves, vol. 25, no. 2, pp. 413–426, Feb. 2004. [90] J. Zhao, “Improvement of MMW irradiation uniformity in culture dishes for experiments on MMW biological effects,” Microw. Opt. Technol. Lett., vol. 40, no. 3, pp. 258–261, Feb. 2004. [91] A. V. Boriskin et al., “Enhancing exposure efficiency and uniformity using a choke ring antenna: application to bioelectromagnetic studies at 60 GHz,” IEEE Trans. Microw. Theory Tech., vol. 61, no. 5, pp. 2005–2014, May 2013. [92] A. G. Pakhomov, Y. Akyel, O. N. Pakhomova, B. E. Stuck, and M. R. Murphy, “Current state and implications of research on biological effects of millimeter waves: A review of the literature,” Bioelectromagnetics, vol. 19, no. 7, pp. 393–413, 1998. [93] A. Ramundo-Orlando, “Effects of millimeter waves radiation on cell membrane - A brief review,” J. Infrared Millim. Terahertz Waves, vol. 31, no. 12, pp. 1400–1411, Dec. 2010.

Bibliography

[94] Y. Le Dréan et al., “State of knowledge on biological effects at 40–60 GHz,” Comptes Rendus Phys., vol. 14, no. 5, pp. 402–411, May 2013. [95] O. V. Betskii and N. N. Lebedeva, “Low-intensity millimeter waves in biology and medicine,” Crit. Rev. Biomed. Eng., vol. 28, no. (1-2), pp. 247–68, Feb. 2000. [96] S. I. Alekseev, M. C. Ziskin, N. V. Kochetkova, and M. A. Bolshakov, “Millimeter waves thermally alter the firing rate of the Lymnaea pacemaker neuron,” Bioelectromagnetics, vol. 18, no. 2, pp. 89–98, Jan. 1997. [97] S. Romanenko, P. H. Siegel, D. A. Wagenaar, and V. Pikov, “Effects of millimeter wave irradiation and equivalent thermal heating on the activity of individual neurons in the leech ganglion,” J. Neurophysiol., vol. 112, no. 10, pp. 2423–2431, Nov. 2014. [98] A. J. Haas, Y. Le Page, M. Zhadobov, R. Sauleau, and Y. Le Dréan, “Effects of 60-GHz millimeter waves on neurite outgrowth in PC12 cells using high-content screening,” Neurosci. Lett., vol. 618, pp. 58–65, Apr. 2016. [99] A. Samsonov and S. V. Popov, “The effect of a 94 GHz electromagnetic field on neuronal microtubules,” Bioelectromagnetics, vol. 34, no. 2, pp. 133–144, Feb. 2013. [100] A. Beneduci, G. Chidichimo, R. De Rose, L. Filippelli, S. V. Straface, and S. Venuta, “Frequency and irradiation time-dependant antiproliferative effect of low-power millimeter waves on RPMI 7932 human melanoma cell line,” Anticancer Res., vol. 25, no. 2A, pp. 1023– 1028, Apr. 2005. [101] G. Chidichimo et al., “Selective inhibition of tumoral cells growth by low power millimeter waves,” Anticancer Res, vol. 22, no. 3, pp. 1681–8., Jun. 2002. [102] A. Beneduci, “Evaluation of the potential in vitro antiproliferative effects of millimeter waves at some therapeutic frequencies on RPMI 7932 human skin malignant melanoma cells,” Cell Biochem. Biophys., vol. 55, no. 1, pp. 25–32, 2009. [103] X. Liu, “Millimeter wave treatment promotes chondrocyte proliferation via G1/S cell cycle transition,” Int. J. Mol. Med., Feb. 2012. [104] S. Koyama et al., “Effects of long-term exposure to 60 GHz millimeter-wavelength radiation on the genotoxicity and heat shock protein (Hsp) expression of cells derived from human eye,” Int. J. Environ. Res. Public. Health, vol. 13, no. 8, Aug. 2016. [105] M. K. Logani, A. Bhanushali, M. C. Ziskin, and T. J. Prihoda, “Micronuclei in peripheral blood and bone marrow cells of mice exposed to 42 GHz electromagnetic millimeter waves,” Radiat. Res., vol. 161, no. 3, pp. 341–345, Mar. 2004. [106] M. Zhadobov et al., “Low-power millimeter wave radiations do not alter stress-sensitive gene expression of chaperone proteins,” Bioelectromagnetics, vol. 28, no. 3, pp. 188–196, Apr. 2007. [107] R. I. Morimoto, “The heat shock response: systems biology of proteotoxic stress in aging and disease,” Cold Spring Harb. Symp. Quant. Biol., vol. 76, pp. 91–99, 2011. [108] D. Habauzit et al., “Transcriptome analysis reveals the contribution of thermal and the specific effects in cellular response to millimeter wave exposure,” PLoS ONE, vol. 9, no. 10, p. e109435, Oct. 2014. [109] C. Nicolas Nicolaz et al., “Absence of direct effect of low-power millimeter-wave radiation at 60.4 GHz on endoplasmic reticulum stress,” Cell Biol. Toxicol., vol. 25, no. 5, pp. 471–478, Oct. 2009. [110] C. N. Nicolaz et al., “Study of narrow band millimeter-wave potential interactions with endoplasmic reticulum stress sensor genes,” Bioelectromagnetics, vol. 30, no. 5, pp. 365–373, 2009. [111] Y. Soubere Mahamoud et al., “Additive effects of millimeter waves and 2-deoxyglucose co- exposure on the human keratinocyte transcriptome,” PLOS ONE, vol. 11, no. 8, p. e0160810, Aug. 2016.

Bibliography

[112] H. Fröhlich, “Theoretical physics and biology,” in Biological Coherence and Response to External Stimuli, H. Fröhlich, Ed. Berlin, Heidelberg: Springer Berlin Heidelberg, 1988, pp. 1– 24. [113] H. Frohlich, “Long-range coherence and energy storage in biological systems,” Int. J. Quantum Chem., vol. 2, pp. 641–649, 1968. [114] A. Ramundo-Orlando et al., “The response of giant phospholipid vesicles to millimeter waves radiation,” Biochim. Biophys. Acta BBA - Biomembr., vol. 1788, no. 7, pp. 1497–1507, Jul. 2009. [115] A. Beneduci, K. Cosentino, and G. Chidichimo, “Millimeter wave radiations affect membrane hydration in phosphatidylcholine vesicles,” Materials, vol. 6, no. 7, pp. 2701–2712, Jul. 2013. [116] S. I. Alekseev and M. C. Ziskin, “Millimeter microwave effect on ion transport across lipid bilayer membranes,” Bioelectromagnetics, vol. 16, no. 2, pp. 124–131, 1995. [117] C. F. Blackman, C. M. Elder, C. M. Weil, S. G. Benane, D. C. Eichinger, and D. E. House, “Induction of calcium-ion efflux from brain tissue by radio-frequency radiation: effects of modulation frequency and field strength,” Radio Sci., vol. 14, no. 6S, pp. 93–98, 1979. [118] I. Szabo, J. Kappelmayer, S. I. Alekseev, and M. C. Ziskin, “Millimeter wave induced reversible externalization of phosphatidylserine molecules in cells exposed in vitro,” Bioelectromagnetics, vol. 27, no. 3, pp. 233–244, Apr. 2006. [119] M. G. Shapiro, M. F. Priest, P. H. Siegel, and F. Bezanilla, “Thermal mechanisms of millimeter wave stimulation of excitable cells,” Biophys. J., vol. 104, no. 12, pp. 2622–2628, Jun. 2013. [120] M. Zhadobov, R. Sauleau, Y. Le Drean, S. I. Alekseev, and M. C. Ziskin, “Numerical and experimental millimeter-wave dosimetry for in vitro experiments,” IEEE Trans. Microw. Theory Tech., vol. 56, no. 12, pp. 2998–3007, Dec. 2008. [121] A. Paffi, M. Liberti, F. Apollonio, A. Sheppard, and Q. Balzano, “In vitro exposure: Linear and non-linear thermodynamic events in Petri dishes,” Bioelectromagnetics, vol. 36, no. 7, pp. 527–537, Oct. 2015. [122] J. X. Zhao, “Numerical dosimetry for cells under millimetre-wave irradiation using Petri dish exposure set-ups,” Phys. Med. Biol., vol. 50, no. 14, p. 3405, 2005. [123] J. Zhao, “In Vitro dosimetry and temperature evaluations of a typical millimeter-wave aperture-field exposure setup,” IEEE Trans. Microw. Theory Tech., vol. 60, no. 11, pp. 3608– 3622, Nov. 2012. [124] A. J. Haas, Y. Le Page, M. Zhadobov, R. Sauleau, Y. L. Dréan, and C. Saligaut, “Effect of acute millimeter wave exposure on dopamine metabolism of NGF-treated PC12 cells,” J. Radiat. Res. (Tokyo), vol. 58, no. 4, pp. 439–445, Jul. 2017. [125] T. Schmid, O. Egger, and N. Kuster, “Automated E-field scanning system for dosimetric assessments,” IEEE Trans. Microw. Theory Tech., vol. 44, no. 1, pp. 105–113, Jan. 1996. [126] S. I. Alekseev and M. C. Ziskin, “Local heating of human skin by millimeter waves: a kinetics study,” Bioelectromagnetics, vol. 24, no. 8, pp. 571–581, Dec. 2003. [127] S. I. Alekseev, M. C. Ziskin, and E. E. Fesenko, “Problems of using a thermocouple for measurements of skin temperature rise during the exposure to millimeter waves,” Biophysics, vol. 56, no. 3, p. 525, Aug. 2011. [128] National Instruments, “Fundamentals of Fiber Bragg Grating (FBG) Optical sensing.” [Online]. Available: https://www.element14.com/community/docs/DOC-87581/l/fundamentals- of-fiber-bragg-grating-fbg-optical-sensing-national-instruments. [129] É. Pinet, S. Ellyson, and F. Borne, “Temperature fiber-optic point sensors: Commercial technologies and industrial applications,” p. 13, 2010. [130] E. Schena, D. Tosi, P. Saccomandi, E. Lewis, and T. Kim, “Fiber optic sensors for temperature monitoring during thermal treatments: an overview,” Sensors, vol. 16, no. 7, p. 1144, Jul. 2016. 73

Bibliography

[131] M. Zhadobov et al., “Exposure system and dosimetry for in vitro studies of biocompatibility of pulse-modulated RF signals of ultrahigh field MRI,” IEEE Trans. Biomed. Eng., vol. 60, no. 11, pp. 3167–3175, Nov. 2013. [132] M. Zhadobov, S. I. Alekseev, R. Sauleau, Y. Le Page, Y. Le Dréan, and E. E. Fesenko, “Microscale temperature and SAR measurements in cell monolayer models exposed to millimeter waves,” Bioelectromagnetics, vol. 38, no. 1, pp. 11–21, Jan. 2017. [133] S. I. Alekseev, M. V. Zhadobov, E. E. Fesenko, and E. E. Fesenko, “Millimeter wave dosimetry at exposure of cell monolayers,” Biophysics, vol. 62, no. 2, pp. 261–264, Mar. 2017. [134] L. Alon, S. Gabriel, G. Y. Cho, R. Brown, and C. M. Deniz, “Prospects for millimeter-wave compliance measurement technologies,” IEEE Antennas Propag. Mag., vol. 59, no. 2, pp. 115– 125, Apr. 2017. [135] M. K. Logani and M. C. Ziskin, “Millimeter waves at 25 mw/cm2 have no effect on hydroxyl radical-dependent lipid peroxidation,” Electro- Magnetobiology, vol. 17, no. 1, pp. 67–73, Jan. 1998. [136] Hallcrest, LCR, Glenview, “Handbook of Thermochromic Liquid Crystal Technology,” 1991. [Online]. Available: http://www.oalib.com/references/13107200. [137] J. L. Fergason, “Liquid crystals in nondestructive testing,” Appl. Opt., vol. 7, no. 9, p. 1729, Sep. 1968. [138] I. Sage, “Thermochromic liquid crystals,” Liq. Cryst., vol. 38, no. 11–12, pp. 1551–1561, Nov. 2011. [139] D. Dabiri, “Digital particle image thermometry/velocimetry: a review,” Exp. Fluids, vol. 46, no. 2, pp. 191–241, Feb. 2009. [140] D. Dabiri and M. Gharib, “Digital particle image thermometry: The method and implementation,” Exp. Fluids, vol. 11, no. 2, pp. 77–86, May 1991. [141] N. Abdullah, Abd. R. Abu Talib, A. A. Jaafar, M. A. Mohd Salleh, and W. T. Chong, “The basics and issues of thermochromic liquid crystal calibrations,” Exp. Therm. Fluid Sci., vol. 34, no. 8, pp. 1089–1121, Nov. 2010. [142] LCR Hallcrest, “TLC products for use in research and testing applications,” p. 18. [143] M. Kojima et al., “Characteristics of ocular temperature elevations after exposure to quasi- and millimeter waves (18-40 GHz),” J. Infrared Millim. Terahertz Waves, vol. 36, no. 4, pp. 390–399, Apr. 2015. [144] M. Kojima et al., “Ocular response to millimeter wave exposure under different levels of humidity,” J. Infrared Millim. Terahertz Waves, vol. 40, no. 5, pp. 574–584, May 2019. [145] M. Kojima et al., “Investigation of acute ocular injury threshold by 76 GHz band exposure in rabbits,” in 2011 XXXth URSI General Assembly and Scientific Symposium, 2011, pp. 1–4. [146] D. W. McRobbie, E. A. Moore, M. J. Graves, and M. R. Prince, MRI from picture to proton. Cambridge University Press, 2007. [147] B. Denis de Senneville, B. Quesson, and C. T. W. Moonen, “Magnetic resonance temperature imaging,” Int. J. Hyperth. Off. J. Eur. Soc. Hyperthermic Oncol. North Am. Hyperth. Group, vol. 21, no. 6, pp. 515–531, Sep. 2005. [148] L. Alon et al., “mmWave exposure assessment using magnetic resonance thermal imaging,” presented at the Bioelectromagnetics Society Annual Meeting, Asilomar Conference Center, California, 2015. [149] Alon Leeor, “mmWave Exposure Assessment using Magnetic Resonance Thermal Imaging.” [150] L. Alon, G. Y. Cho, X. Yang, D. K. Sodickson, and C. M. Deniz, “A method for safety testing of radiofrequency/microwave-emitting devices using MRI,” Magn. Reson. Med., vol. 74, no. 5, pp. 1397–1405, 2015. [151] L. Alon, D. K. Sodickson, and C. M. Deniz, “Heat equation inversion framework for average SAR calculation from magnetic resonance thermal imaging,” Bioelectromagnetics, vol. 37, no. 7, pp. 493–503, Oct. 2016. 74

Bibliography

[152] I. Mellal, A. Oukaira, E. Kengene, and A. Lakhssassi, “Thermal therapy modalities for cancer treatment: a review and future perspectives,” Int. J. Appl. Sci. - Res. Rev., vol. 4, no. 4, Nov. 2017. [153] J. H. Breasted, “The Edwin Smith surgical papyrus, volume 1: hieroglyphic transliteration, translation, and commentary,” Oriental Institute Publications (OIP), University of Chicago. [154] N. B. Hornback, “Historical aspects of hyperthermia in cancer therapy,” Radiol. Clin. North Am., vol. 27, no. 3, pp. 481–488, May 1989. [155] R. W. Y. Habash, R. Bansal, D. Krewski, and H. T. Alhafid, “Thermal Therapy, Part 1: an introduction to thermal therapy,” Crit. Rev. Biomed. Eng., vol. 34, no. 6, pp. 459–489, 2006. [156] W. Busch, “Uber den einfluss welche heftigere erysipeln zuweilig auf organisierte neubildungenausuben,” Verhandlungen des Naturhistorischen Vereines der Preussischen Rheinlande undWestphalens, vol. 23, no. 28–30, 1886. [157] E. F. McCarthy, “The toxins of William B. Coley and the treatment of bone and soft-tissue sarcomas,” Iowa Orthop. J., vol. 26, pp. 154–158, 2006. [158] J. R. Stewart, “Prospects for hyperthermia in cancer treatment,” in Hyperthermia and oncology, vol. 1, VSP, Utrecht, 1988, pp. 1–12. [159] S. Roussakow, “The history of hyperthermia rise and decline,” Conf. Pap. Med., vol. 2013, pp. 1–40, 2013. [160] W. C. Dewey, “Arrhenius relationships from the molecule and cell to the clinic,” Int. J. Hyperth. Off. J. Eur. Soc. Hyperthermic Oncol. North Am. Hyperth. Group, vol. 25, no. 1, pp. 3–20, Feb. 2009. [161] G. Crile, “Heat as an adjunct to the treatment of cancer; experimental studies,” Cleve. Clin. Q., vol. 28, pp. 75–89, Apr. 1961. [162] G. Crile, “Selective destruction of cancers after exposure to heat,” Ann. Surg., vol. 156, no. 3, pp. 404–407, Sep. 1962. [163] G. Crile, “The effects of heat and radiation on cancers implanted on the feet of mice,” Cancer Res., vol. 23, p. 10, 1963. [164] Z. Behrouzkia, Z. Joveini, B. Keshavarzi, N. Eyvazzadeh, and R. Z. Aghdam, “Hyperthermia: how can it be used?,” Oman Med. J., vol. 31, no. 2, pp. 89–97, Mar. 2016. [165] P. Wust et al., “Hyperthermia in combined treatment of cancer,” Lancet Oncol., vol. 3, no. 8, pp. 487–497, Aug. 2002. [166] E. Cabuy, “Hyperthermia in cancer treatment,” Reliable cancer therapies. Energy-based therapies., vol. 1, no. 2, pp. 1–48, 2011. [167] L. G. P. H. Vroomen, E. N. Petre, F. H. Cornelis, S. B. Solomon, and G. Srimathveeravalli, “Irreversible electroporation and thermal ablation of tumors in the liver, lung, kidney and bone: What are the differences?,” Diagn. Interv. Imaging, vol. 98, no. 9, pp. 609–617, Sep. 2017. [168] G. S. Gazelle, S. N. Goldberg, L. Solbiati, and T. Livraghi, “Tumor ablation with radiofrequency energy,” Radiology, vol. 217, no. 3, pp. 633–646, Dec. 2000. [169] M. G. Lubner, C. L. Brace, J. L. Hinshaw, and F. T. Lee, “Microwave tumor ablation: mechanism of action, clinical results and devices,” J. Vasc. Interv. Radiol. JVIR, vol. 21, no. 8 Suppl, pp. S192–S203, Aug. 2010. [170] Z. Izadifar, P. Babyn, and D. Chapman, “Mechanical and biological effects of ultrasound: a review of present knowledge,” Ultrasound Med. Biol., vol. 43, no. 6, pp. 1085–1104, Jun. 2017. [171] O. Al-Bataineh, J. Jenne, and P. Huber, “Clinical and future applications of high intensity focused ultrasound in cancer,” Cancer Treat. Rev., vol. 38, no. 5, pp. 346–353, Aug. 2012. [172] J. Tu et al., “Controllable in vivo hyperthermia effect induced by pulsed high intensity focused ultrasound with low duty cycles,” Appl. Phys. Lett., vol. 101, no. 12, p. 124102, Sep. 2012. [173] J. Yu et al., “Self-assembly synthesis, tumor cell targeting, and photothermal capabilities of antibody-coated indocyanine green nanocapsules,” J. Am. Chem. Soc., vol. 132, no. 6, pp. 1929–1938, Feb. 2010. 75

Bibliography

[174] D. Li, J. Huang, and R. B. Kaner, “Polyaniline nanofibers: a unique polymer nanostructure for versatile applications,” Acc. Chem. Res., vol. 42, no. 1, pp. 135–145, Jan. 2009. [175] M. Li, X. Yang, J. Ren, K. Qu, and X. Qu, “Using graphene oxide high near-infrared absorbance for photothermal treatment of Alzheimer’s disease,” Adv. Mater. Deerfield Beach Fla, vol. 24, no. 13, pp. 1722–1728, Apr. 2012. [176] Z. Bao, X. Liu, Y. Liu, H. Liu, and K. Zhao, “Near-infrared light-responsive inorganic nanomaterials for photothermal therapy,” Asian J. Pharm. Sci., vol. 11, no. 3, pp. 349–364, Jun. 2016. [177] L. Zou et al., “Current approaches of photothermal therapy in treating cancer metastasis with nanotherapeutics,” Theranostics, vol. 6, no. 6, pp. 762–772, 2016. [178] R. W. Y. Habash, R. Bansal, D. Krewski, and H. T. Alhafid, “Thermal therapy, part III: ablation techniques,” Crit. Rev. Biomed. Eng., vol. 35, no. 1–2, 2007. [179] Y. Shi and Y. Mao, “Magnetic resonance thermometry–guided laser interstitial thermal therapy in neurosurgery, a promising tool for dural-based lesions?,” World Neurosurg., vol. 98, pp. 836–838, Feb. 2017. [180] M. E. Khosroshahi, Z. Hassannejad, M. Firouzi, and A. R. Arshi, “Nanoshell-mediated targeted photothermal therapy of HER2 human breast cancer cells using pulsed and continuous wave lasers: an in vitro study,” Lasers Med. Sci., vol. 30, no. 7, pp. 1913–1922, Sep. 2015. [181] M. Soncin, A. Busetti, F. Fusi, G. Jori, and M. A. Rodgers, “Irradiation of amelanotic melanoma cells with 532 nm high peak power pulsed laser radiation in the presence of the photothermal sensitizer Cu(II)-hematoporphyrin: a new approach to cell photoinactivation,” Photochem. Photobiol., vol. 69, no. 6, pp. 708–712, Jun. 1999. [182] J. Patterson and R. Strang, “The role of blood flow in hyperthermia,” Int. J. Radiat. Oncol. • Biol. • Phys., vol. 5, no. 2, pp. 235–241, Feb. 1979. [183] M. Gautherie, Ed., Biological basis of oncologic thermotherapy. Berlin Heidelberg: Springer-Verlag, 1990. [184] S. I. Masunaga, “Tumor microenvironment and hyperthermia,” in Hyperthermic Oncology from Bench to Bedside, Springer, 2016. [185] P. Vaupel, F. Kallinowski, and P. Okunieff, “Blood flow, oxygen and nutrient supply, and metabolic microenvironment of human tumors: a review,” Cancer Res., vol. 49, no. 23, pp. 6449–6465, Dec. 1989. [186] B. Hildebrandt et al., “The cellular and molecular basis of hyperthermia,” Crit. Rev. Oncol. Hematol., vol. 43, no. 1, pp. 33–56, Jul. 2002. [187] J. R. Lepock, “How do cells respond to their thermal environment?,” Int. J. Hyperthermia, vol. 21, no. 8, pp. 681–687, Dec. 2005. [188] J. L. Roti Roti, “Cellular responses to hyperthermia (40-46 degrees C): cell killing and molecular events,” Int. J. Hyperth. Off. J. Eur. Soc. Hyperthermic Oncol. North Am. Hyperth. Group, vol. 24, no. 1, pp. 3–15, Feb. 2008. [189] A. Laszlo, “The effects of hyperthermia on mammalian cell structure and function,” Cell Prplif, vol. 25, no. 2, pp. 59–87, Mar. 1992. [190] H. H. Kampinga, J. R. Dynlacht, and E. Dikomey, “Mechanism of radiosensitization by hyperthermia (> or = 43 degrees C) as derived from studies with DNA repair defective mutant cell lines,” Int. J. Hyperth. Off. J. Eur. Soc. Hyperthermic Oncol. North Am. Hyperth. Group, vol. 20, no. 2, pp. 131–139, Mar. 2004. [191] G. C. Li, N. F. Mivechi, and G. Weitzel, “Heat shock proteins, thermotolerance, and their relevance to clinical hyperthermia,” Int. J. Hyperthermia, vol. 11, no. 4, pp. 459–488, Jan. 1995. [192] J. R. Lepock, “Protein denaturation during heat shock,” in Advances in Molecular and Cell Biology, vol. 19, Elsevier, 1997, pp. 223–259.

Bibliography

[193] J. R. Lepock, “Cellular effects of hyperthermia: relevance to the minimum dose for thermal damage,” Int. J. Hyperth. Off. J. Eur. Soc. Hyperthermic Oncol. North Am. Hyperth. Group, vol. 19, no. 3, pp. 252–266, Jun. 2003. [194] J. R. Lepock, “Role of nuclear protein denaturation and aggregation in thermal radiosensitization,” Int. J. Hyperth. Off. J. Eur. Soc. Hyperthermic Oncol. North Am. Hyperth. Group, vol. 20, no. 2, pp. 115–130, Mar. 2004. [195] R. VanderWaal, R. S. Malyapa, R. Higashikubo, and J. L. Roti Roti, “A comparison of the modes and kinetics of heat-induced cell killing in HeLa and L5178Y cells,” Radiat. Res., vol. 148, no. 5, pp. 455–62, Nov. 1997. [196] P. Syntichaki and N. Tavernarakis, “Death by necrosis,” EMBO Rep., vol. 3, no. 7, pp. 604– 609, Jul. 2002. [197] H. Vakifahmetoglu, M. Olsson, and B. Zhivotovsky, “Death through a tragedy: mitotic catastrophe,” Cell Death Differ., vol. 15, no. 7, pp. 1153–1162, Jul. 2008. [198] P. Schildkopf et al., “Biological rationales and clinical applications of temperature controlled hyperthermia - implications for multimodal cancer treatments,” Curr. Med. Chem., vol. 17, no. 27, pp. 3045–3057, Sep. 2010. [199] A. Takahashi, “Inhibition of DNA repair system activity,” in Hyperthermic oncology from bench to bedside, Springer, 2016. [200] G. Iliakis, W. Wu, and M. Wang, “DNA double strand break repair inhibition as a cause of heat radiosensitization: Re-evaluation considering backup pathways of NHEJ,” Int. J. Hyperthermia, vol. 24, no. 1, pp. 17–29, Jan. 2008. [201] M. R. Horsman and J. Overgaard, “Hyperthermia: a potent enhancer of radiotherapy,” Clin. Oncol., vol. 19, no. 6, pp. 418–426, Aug. 2007. [202] C. W. Song, “Effect of local hyperthermia on blood flow and microenvironment: a review,” Cancer Res., vol. 44, no. 10 Suppl, pp. 4721s–4730s, Oct. 1984. [203] H. S. Reinhold, “Physiological effects of hyperthermia,” in Application of hyperthermia in the treatment of cancer, 1988, pp. 32–43. [204] R. E. Durand, “Keynote address: the influence of microenvironmental factors on the activity of radiation and drugs,” Int. J. Radiat. Oncol. Biol. Phys., vol. 20, no. 2, pp. 253–258, Feb. 1991. [205] M. Mallory, E. Gogineni, G. C. Jones, L. Greer, and C. B. Simone, “Therapeutic hyperthermia: The old, the new, and the upcoming,” Crit. Rev. Oncol. Hematol., vol. 97, pp. 56–64, Jan. 2016. [206] T. Ohtsubo et al., “Enhancement of cisplatin sensitivity and platinum uptake by 40 degrees C hyperthermia in resistant cells,” Cancer Lett., vol. 119, no. 1, pp. 47–52, Oct. 1997. [207] L. Li et al., “Mild hyperthermia triggered doxorubicin release from optimized stealth thermosensitive liposomes improves intratumoral drug delivery and efficacy,” J. Controlled Release, vol. 168, no. 2, pp. 142–150, Jun. 2013. [208] C. D. Landon, S. E. Benjamin, K. A. Ashcraft, and M. W. Dewhirst, “A role for the copper transporter Ctr1 in the synergistic interaction between hyperthermia and cisplatin treatment,” Int. J. Hyperthermia, vol. 29, no. 6, pp. 528–38, 2013. [209] M. N. Rylander, “Thermally induced injury and heat-shock protein expression in cells and tissues,” Ann. N. Y. Acad. Sci., vol. 1066, no. 1, pp. 222–242, Dec. 2005. [210] G. M. Hahn and G. C. Li, “Thermotolerance and heat shock proteins in mammalian cells,” Radiat. Res., vol. 92, no. 3, pp. 452–457, Dec. 1982. [211] C. Jolly, “Role of the heat shock response and molecular chaperones in oncogenesis and cell death,” J. Natl. Cancer Inst., vol. 92, no. 19, pp. 1564–1572, Oct. 2000. [212] C. Wu, “Heat shock transcription factors: structure and regulation,” Annu. Rev. Cell Dev. Biol., vol. 11, pp. 441–469, 1995. [213] R. A. Krebs and M. E. Feder, “Deleterious consequences of Hsp70 overexpression in Drosophila melanogaster larvae,” Cell Stress Chaperones, vol. 2, no. 1, pp. 60–71, Mar. 1997. 77

Bibliography

[214] J. E. Podrabsky and G. N. Somero, “An inducible 70 kDa-class heat shock protein is constitutively expressed during early development and diapause in the annual killifish Austrofundulus limnaeus,” Cell Stress Chaperones, vol. 12, no. 3, pp. 199–204, Aug. 2007. [215] A.-P. Arrigo, “Analysis of HspB1 (Hsp27) oligomerization and phosphorylation patterns and its interaction with specific client polypeptides,” Methods Mol. Biol. Clifton NJ, vol. 1709, pp. 163–178, 2018. [216] S. B. Field, “The concept of thermal dose,” Recent Results Cancer Res. Fortschritte Krebsforsch. Progres Dans Rech. Sur Cancer, vol. 107, pp. 1–6, 1988. [217] R. T. Pettigrew, J. M. Galt, C. M. Ludgate, D. B. Horn, and A. N. Smith, “Circulatory and biochemical effects of whole body hyperthermia,” Br. J. Surg., vol. 61, no. 9, pp. 727–730, Sep. 1974. [218] Y. Okumura and H. S. Reinhold, “Heat sensitivity of rat skin,” Eur. J. Cancer 1965, vol. 14, no. 11, pp. 1161–1166, Nov. 1978. [219] J. E. Robinson, M. J. Wizemberg, and W. McCready, “Radiation and hyperthermal response of normal tissue in situ,” Radiology, vol. 113, no. 1, pp. 195–8, 1974. [220] R. T. Pettigrew, J. M. Galt, C. M. Ludgate, and A. N. Smith, “Clinical effects of whole-body hyperthermia in advanced malignancy,” BMJ, vol. 4, no. 5946, pp. 679–682, Dec. 1974. [221] J. S. Stehlin, B. C. Giovanella, P. D. de Ipolyi, L. R. Muenz, and R. F. Anderson, “Results of hyperthermic perfusion for melanoma of the extremities,” Surg. Gynecol. Obstet., vol. 140, no. 3, pp. 339–348, Mar. 1975. [222] J. Mendecki, E. Friedenthal, and C. Botstein, “Effects of microwave-induced local hyperthermia on mammary adenocarcinoma in C3H mice,” Cancer Res., vol. 36, no. 6, pp. 2113–4, Jul. 1976. [223] H. H. LeVeen, S. Wapnick, V. Piccone, G. Falks, and N. Ahmed, “Tumor eradication by radiofrequency therapy: response in 21 patients,” AMA, vol. 17, no. 235(20), pp. 2198–200. [224] K. J. Henle and L. A. Dethlefsen, “Time-temperature relationships for heat-induced killing of mammalian cells,” Ann. N. Y. Acad. Sci., vol. 335, no. 1, pp. 234–253, 1980. [225] E. R. Atkinson, “Hyperthermia dose definition,” in MTT-S International Microwave Symposium Digest, San Diego, CA, USA, 1977, vol. 77, pp. 251–251. [226] E. W. Gemer, “Thermal dose: definition and evaluation in terms of hyperthermia induced cytotoxicity,” presented at the 7th Int. Congress of Radiat. Res., Amsterdam, the Netherlands, 1983. [227] G. M. Hahn, Hyperthermia and cancer. Springer US, 1982. [228] S. B. Field and C. C. Morris, “The relationship between heating time and temperature: its relevance to clinical hyperthermia,” Radiother. Oncol., vol. 1, no. 2, pp. 179–186, Nov. 1983. [229] S. A. Sapareto, “Thermal isoeffect dose: Addressing the problem of thermotolerance,” Int. J. Hyperthermia, vol. 3, no. 4, pp. 297–305, Jan. 1987. [230] G. C. van Rhoon, “Is CEM43 still a relevant thermal dose parameter for hyperthermia treatment monitoring?,” Int. J. Hyperth. Off. J. Eur. Soc. Hyperthermic Oncol. North Am. Hyperth. Group, vol. 32, no. 1, pp. 50–62, 2016. [231] S. Arrhenius, “Über die Reaktionsgeschwindigkeit bei der Inversion von Rohrzucker durch Säuren,” Z. Für Phys. Chem., Jan. 1889. [232] J. A. Pearce, “Comparative analysis of mathematical models of cell death and thermal damage processes,” Int. J. Hyperthermia, vol. 29, no. 4, pp. 262–280, Jun. 2013. [233] X. He and J. C. Bischof, “Quantification of temperature and injury response in thermal therapy and cryosurgery,” Crit. Rev. Biomed. Eng., vol. 31, no. 5–6, pp. 355–422, 2003. [234] J. A. Pearce, “Thermal dose models: irreversible alteration in tissues.,” in Physics of thermal therapy: fundamentals and clinical applications, First Edn., CRC Press, 2012. [235] X. He, S. Bhowmick, and J. C. Bischof, “Thermal therapy in urologic systems: a comparison of arrhenius and thermal isoeffective dose models in predicting hyperthermic injury,” J. Biomech. Eng., vol. 131, no. 7, p. 074507, Jul. 2009. 78

Bibliography

[236] F. C. Henriques, “Studies of thermal injury; the predictability and the significance of thermally induced rate processes leading to irreversible epidermal injury,” Arch. Pathol., vol. 43, no. 5, pp. 489–502, May 1947. [237] F. C. Henriques and A. R. Moritz, “Studies of thermal injury: I. The conduction of heat to and through skin and the temperatures attained therein. A theoretical and an experimental investigation,” Am. J. Pathol., vol. 23, no. 4, pp. 530–549, Jul. 1947. [238] A. R. Moritz and F. C. Henriques, “Studies of thermal Injury: II. The relative importance of time and surface temperature in the causation of cutaneous burns,” Am. J. Pathol., vol. 23, no. 5, pp. 695–720, Sep. 1947. [239] A. R. Moritz, “Studies of thermal injury: III. The pathology and pathogenesis of cutaneous burns. An experimental study,” Am. J. Pathol., vol. 23, no. 6, pp. 915–941, Nov. 1947. [240] S. A. Sapareto and W. C. Dewey, “Thermal dose determination in cancer therapy,” Int. J. Radiat. Oncol. Biol. Phys., vol. 10, no. 6, pp. 787–800, 1984. [241] M. W. Dewhirst, B. L. Viglianti, M. Lora-Michiels, M. Hanson, and P. J. Hoopes, “Basic principles of thermal dosimetry and thermal thresholds for tissue damage from hyperthermia,” Int. J. Hyperth. Off. J. Eur. Soc. Hyperthermic Oncol. North Am. Hyperth. Group, vol. 19, no. 3, pp. 267–294, Jun. 2003. [242] L. Roizin-Towle and J. P. Pirro, “The response of human and rodent cells to hyperthermia,” Int. J. Radiat. Oncol., vol. 20, no. 4, pp. 751–756, Apr. 1991. [243] P. S. Yarmolenko et al., “Thresholds for thermal damage to normal tissues: An update,” Int. J. Hyperth. Off. J. Eur. Soc. Hyperthermic Oncol. North Am. Hyperth. Group, vol. 27, no. 4, pp. 320–343, 2011. [244] M. W. Dewhirst, B. L. Viglianti, M. Lora-Michiels, P. J. Hoopes, and M. Hanson, “Thermal dose requirement for tissue effect: experimental and clinical finding,” Proc. SPIE-- Int. Soc. Opt. Eng., vol. 4954, p. 37, Jun. 2003. [245] M. Breton and L. M. Mir, “Microsecond and nanosecond electric pulses in cancer treatments,” Bioelectromagnetics, vol. 33, no. 2, pp. 106–123, Feb. 2012. [246] M. B. Sano, C. B. Arena, M. R. DeWitt, D. Saur, and R. V. Davalos, “In-vitro bipolar nano- and microsecond electro-pulse bursts for irreversible electroporation therapies,” Bioelectrochemistry, vol. 100, pp. 69–79, Dec. 2014. [247] P. Leveque, R. O’Connor, and D. Arnaud-Cormos, “Measurement and characterization of exposure systems for high-frequency, ultrashort pulses,” in Handbook of Electroporation, D. Miklavčič, Ed. Cham: Springer International Publishing, 2017, pp. 813–836. [248] D. Arnaud-Cormos, V. Couderc, and P. Leveque, “Photoconductive switching for pulsed high-voltage generators,” in Handbook of Electroporation, D. Miklavcic, Ed. Cham: Springer International Publishing, 2017, pp. 1–21. [249] M. L. Yarmush, A. Golberg, G. Serša, T. Kotnik, and D. Miklavčič, “Electroporation-based technologies for medicine: principles, applications, and challenges,” Annu. Rev. Biomed. Eng., vol. 16, no. 1, pp. 295–320, Jul. 2014. [250] N. Esmaeili and M. Friebe, “Electrochemotherapy: a review of current status, alternative IGP approaches, and future perspectives,” J. Healthc. Eng., vol. 2019, pp. 1–11, Jan. 2019. [251] S. J. Beebe, P. M. Fox, L. J. Rec, K. Somers, R. H. Stark, and K. H. Schoenbach, “Nanosecond pulsed electric field (nsPEF) effects on cells and tissues: apoptosis induction and tumor growth inhibition,” IEEE Trans. Plasma Sci., vol. 30, no. 1, pp. 286–292, Feb. 2002. [252] L. Carr, S. M. Bardet, R. C. Burke, D. Arnaud-Cormos, P. Leveque, and R. P. O’Connor, “Calcium-independent disruption of microtubule dynamics by nanosecond pulsed electric fields in U87 human glioblastoma cells,” Sci. Rep., vol. 7, Jan. 2017. [253] R. Nuccitelli et al., “First-in-human trial of nanoelectroablation therapy for basal cell carcinoma: proof of method,” Exp. Dermatol., vol. 23, no. 2, pp. 135–137, Feb. 2014. [254] R. V. Davalos, I. L. M. Mir, and B. Rubinsky, “Tissue ablation with irreversible electroporation,” Ann. Biomed. Eng., vol. 33, no. 2, pp. 223–231, Feb. 2005. 79

Bibliography

[255] K. Schoenbach and S. Xiao, “Method and system for treating a biological target region using pulsed electromagnetic radiation,” WO/2010/151370, 30-Dec-2010. [256] S. L. Jacques, “How tissue optics affect dosimetry for photochemical, photothermal, and photomechanical mechanisms of laser-tissue interaction,” in Recent Advances in the Uses of Light in Physics, Chemistry, Engineering, and Medicine, 1992, vol. 1599, p. 316. [257] M. H. Khan, R. K. Sink, D. Manstein, D. Eimerl, and R. R. Anderson, “Intradermally focused infrared laser pulses: Thermal effects at defined tissue depths,” Lasers Surg. Med., vol. 36, no. 4, pp. 270–280, Apr. 2005. [258] M. Ganguly, “Analyzing thermal and mechanical effects of pulsed laser irradiation on tissues,” Thesis, 2014. [259] K. Mitra and S. Miller, Short pulse laser systems for biomedical applications. Springer International Publishing, 2017. [260] A. Vogel and V. Venugopalan, “Mechanisms of pulsed laser ablation of biological tissues,” Chem. Rev., vol. 103, no. 2, pp. 577–644, Feb. 2003. [261] A. Y. Sajjadi, K. Mitra, and Z. Guo, “Thermal analysis and experiments of laser-tissue interactions: a review,” Heat Transf. Res., vol. 44, no. 3–4, 2013. [262] K.-S. Cheng and R. B. Roemer, “Blood perfusion and thermal conduction effects in Gaussian beam, minimum time single-pulse thermal therapies,” Med. Phys., vol. 32, no. 2, pp. 311–317, Feb. 2005. [263] B. E. Billard, K. Hynynen, and R. B. Roemer, “Effects of physical parameters on high temperature ultrasound hyperthermia,” Ultrasound Med. Biol., vol. 16, no. 4, pp. 409–420, Jan. 1990. [264] R. H. Britt, D. W. Pounds, and B. E. Lyons, “Feasibility of treating malignant brain tumors with focused ultrasound,” in Progress in Tumor Research, vol. 28, M. L. Rosenblum and C. B. Wilson, Eds. S. Karger AG, 1984, pp. 232–245. [265] J. W. Hunt, R. Lalonde, H. Ginsberg, S. Urchuk, and A. Worthington, “Rapid heating: critical theoretical assessment of thermal gradients found in hyperthermia treatments,” Int. J. Hyperth. Off. J. Eur. Soc. Hyperthermic Oncol. North Am. Hyperth. Group, vol. 7, no. 5, pp. 703–718, Oct. 1991. [266] C. Damianou and K. Hynynen, “Focal spacing and near-field heating during pulsed high temperature ultrasound therapy,” Ultrasound Med. Biol., vol. 19, no. 9, pp. 777–787, 1993. [267] K.-S. Cheng and R. B. Roemer, “Closed-form solution for the thermal dose delivered during single pulse thermal therapies,” Int. J. Hyperthermia, vol. 21, no. 3, pp. 215–230, May 2005. [268] K.-S. Cheng and R. B. Roemer, “Optimal power deposition patterns for ideal high temperature therapy/hyperthermia treatments,” Int. J. Hyperthermia, vol. 20, no. 1, pp. 57–72, Feb. 2004. [269] M. Bedoya, A. M. del Rio, J. Chiang, and C. L. Brace, “Microwave ablation energy delivery: Influence of power pulsing on ablation results in an ex vivo and in vivo liver model,” Med. Phys., vol. 41, no. 12, Dec. 2014. [270] E. J. Berjano, J. L. Alió, and J. Saiz, “Modeling for radio-frequency conductive keratoplasty: implications for the maximum temperature reached in the cornea,” Physiol. Meas., vol. 26, no. 3, pp. 157–172, Jun. 2005. [271] M. Ganguly, S. Miller, and K. Mitra, “Model development and experimental validation for analyzing initial transients of irradiation of tissues during thermal therapy using short pulse lasers,” Lasers Surg. Med., vol. 47, no. 9, pp. 711–722, Nov. 2015. [272] B. Zorec, D. Škrabelj, M. Marinček, D. Miklavčič, and N. Pavšelj, “The effect of pulse duration, power and energy of fractional Er:YAG laser for transdermal delivery of differently sized FITC dextrans,” Int. J. Pharm., vol. 516, no. 1–2, pp. 204–213, Jan. 2017. [273] W.-C. Cui et al., “In vivo hyperthermia effect induced by high-intensity pulsed ultrasound,” Chin. Phys. B, vol. 21, no. 7, p. 074301, Jul. 2012.

Chapter 2 Electromagnetic and thermal dosimetry at millimeter waves: effects of thermal convection during in vitro exposure of cells

Content

2.1 Introduction...... 83 2.2 Heat transfer and convection ...... 84

2.2.1 Conduction ...... 84 2.2.2 Radiation ...... 85 2.2.3 Convection...... 87

2.3 Experimental set-up ...... 91 2.4 Samples under test ...... 94

2.4.1 Electromagnetic properties ...... 96 2.4.2 Thermo-physical properties ...... 96

2.5 Temperature measurements ...... 97

2.6 Electromagnetic dosimetry ...... 112

2.6.1 Impact of cell monolayer ...... 120

2.7 Temperature–based retrieval of SAR ...... 122 2.8 Continous wave ...... 126

2.8.1 Heating in different media ...... 126 2.8.2 Dependence on SAR ...... 130 2.8.3 Role of liquid volume ...... 133

2.9 Pulse-induced heating ...... 135

2.9.1 Dependence on pulse duration ...... 135 2.9.2 Heating in water and agar gel ...... 138

2.10 Conclusions ...... 140

Bibliography...... 142

CHAPTER 2. Electromagnetic and thermal dosimetry at millimeter waves: effects of thermal convection during in vitro exposure of cells

2.1 Introduction

Biological effects induced by MMW have been largely studied [1], [2]. As briefly reviewed in the second section of the Chapter 1 (section 1.2.2, Biological effects of millimeter wave) numerous in vitro experiments have been performed during the last 50 years to identify the cellular and molecular mechanisms of interaction between MMW and biological cells. However the outcomes of the biological results are often conflicting and difficult to replicate. The causes of experimental discrepancies are typically attributed to the lack of precise electromagnetic dosimetry and well- controlled exposure conditions. Nevertheless, it has been recently demonstrated [3], [4] that in addition to electromagnetic parameters, thermodynamic events occurring in the culture medium that provides nutrients, hormones, attachment and growing factors to the cells, may be responsible for different biological outcomes under apparently similar exposure conditions (i.e., incident power density). In liquids, the background laminar convection develops under the normal conditions due to a non-uniform temperature distribution near the boundaries of the medium, or natural evaporative cooling [5]. The temperature differences in this case are small resulting only in insignificant steady effect on temperature distribution. In the case of in vitro sample, weak laminar convective currents may result from the lower temperature at the surface of liquid (i.e., cell culture medium) due to its evaporation compared to the temperature at the bottom of the well (typical T is within 1°C), or from the different temperature between the bulk and the walls of the plastic vessel (typically within 0.2°C). The MMW and microwave (MW) power density distribution in in vitro experiments is not uniform [4], [6]–[9] resulting in generation of temperature gradients leading to the formation of density difference in liquid. The latter, drives the natural or free convection flow in liquid directed from the heated bottom of the well to its top with some delay following the formation of a temperature gradient. The surrounding cooler liquid moves to the bottom and after heating reiterates the convection flow [10]–[13]. The initiation of convection in liquids by MMW and MW exposures was described in a number of papers [3], [4], [7], [14]–[16]. Convection during in vitro bioelectromagnetic experiments may greatly impact the temperature dynamics and distribution [17], [18] resulting in small, constant non- Brownian motion of the medium [19], the formation of a toroidal vortex [16], and in gradual temperature drop of the irradiated spot [7], [20]. Besides, liquid motion caused by convective currents may perturb the density of the seeded cell monolayer [21], or the local concentration of oxygen and nutrients carried by the culture medium near the surface of the cells [4]. However,

CHAPTER 2. Electromagnetic and thermal dosimetry at millimeter waves: effects of thermal convection during in vitro exposure of cells convection-induced thermodynamic events are absent in non-exposed sham samples (i.e., changes in biochemical concentrations in sham would be affected only by diffusion), potentially leading to misinterpretation of biological outcome. The existence of convective currents in liquid media inserts doubts about whether electromagnetic energy alone is the cause of biological effects observed in some in vitro experimental studies. For this reason, in order to extend the knowledge about in vitro thermal dosimetry at MMW, we analyzed in detail the effects of the convection occurring in liquid samples exposed to MMW in continuous (CW) and pulsed modulated (PW) regimes. Using several models, including distilled water, culture medium, and agar gel (convection-free water-equivalent model), we investigated the effects of convection on local heating by comparing temperature kinetics locally recorded in the liquids to that in the water-equivalent gel without convection. Convection is studied as a function of several parameters, which may differ from one in vitro experiment to another, including i) viscosity, ii) specific absorption rate (SAR), iii) liquid volume, and iv) duration of a thermal pulse. The chapter is organized as follows. First, theoretical basis of the heat transfer are briefly outlined. Second, material and methods including the samples used and the electromagnetic and thermal dosimetry adopted are described. Third, the temperature dynamics in the exposed samples are presented and analyzed. Finally, the main conclusions are drawn.

2.2 Heat transfer and convection

Heat transfer represents the thermal energy in transit when a spatial temperature difference occurs in the same or different mediums [13]. The transfer of energy is always from the higher- temperature medium to the lower-temperature one. Depending on the molecules arrangement during the heat transfer process, three main modes are defined namely, conduction, irradiation, and convection. Basic theory of heat transfer modes mainly focused on thermal convection is provided below to allow the complete understanding of the problematic addressed in this chapter.

2.2.1 Conduction Conduction heat transfer occurs when a temperature gradient is generated in a solid, liquid or gas in absence of bulk motion. The energy associated to the particles is transferred from the more energetic to the adjacent less energetic ones as the result of their interaction. In liquids and gases, conduction is due both to the collision and the diffusion of the molecules during their random

CHAPTER 2. Electromagnetic and thermal dosimetry at millimeter waves: effects of thermal convection during in vitro exposure of cells interaction. In solids, where the molecules are closely packed into a lattice, conduction is due to the combination of vibrations of the molecules and energy transport by free electrons. The equation governing conduction heat transfer is known as Fourier’s law and it allows to determine the conduction heat transfer rate. For the one-dimensional plane wall shown in Figure 2.1 the equation is expressed as:

(2.1)

The heat flux (W/m²) is the heat transfer rate in the x direction, per unit area perpendicular to the direction of the heat transfer. The heat flux is proportional to the temperature gradient ∂T/∂x, which represents the slope of the temperature curve on a T-x diagram (i.e., the rate of change of the temperature along the x direction) at a certain location x. Heat is conducted in the direction of decreasing temperature, and the temperature gradient becomes negative when temperature decreases with increasing x. The negative sign in Equation (2.1) ensures that heat transfer in the positive x direction is a positive quantity. The constant of proportionality k (W/(m∙°C)) represents the thermal conductivity of the material (i.e., its ability to conduct heat) as described in the paragraph 1.1.2 (section Thermal properties of the skin) of the first chapter of this thesis.

Figure 2.1: 1–D heat transfer by conduction through a large plane wall.

2.2.2 Radiation Radiative heat transfer is the energy emitted by the matter in the form of electromagnetic waves or photons as the results of the changes in the electronic configuration of the molecules. The transfer of energy by radiation does not require the presence of a material medium to travel. In fact, radiation transfer occurs most efficiently in vacuum.

CHAPTER 2. Electromagnetic and thermal dosimetry at millimeter waves: effects of thermal convection during in vitro exposure of cells

The rate of radiation that can be emitted from a surface is expressed by the Stephan-Bolzman law, which states that the total energy radiated by a black body across per unit time is proportional to the fourth power of the blackbody's thermodynamic temperature :

(2.2)

where (W) is the energy emitted by the surface, σ is the Stefan–Boltzmann constant

( ), and is the absolute temperature of the surface (K). This formula is referring to an idealized surface, known as blackbody, which emits radiation at its maximum rate. The heat flux emitted by a real surface is less than that of a blackbody at the same temperature and is given by:

(2.3)

where εB is the emissivity of the surface A whose value is in the range 0 ≤ εB ≤ 1. Radiation may be also incident on a surface from its surroundings ( ). It represents the radiation that may originate from a special source such as the sun or a particular surface of interest. Part of this radiation is absorbed ( ) by the surface as:

(2.4)

where is the absorptivity of the material, which is the fraction of the radiation energy incident on a surface that is absorbed by the surface. Like emissivity, its value is in the range 0 ≤ ≤ 1. A blackbody absorbs the entire radiation incident on it: a blackbody is a perfect absorber ( = 1) as it is a perfect emitter (εB = 1).

In a common situation (Fig. 2.2) where a small surface of emissivity εB and surface area As at the absolute temperature Ts is enclosed by a much larger (or black) surface at absolute temperature Tsurr separated by a gas (such as air) that does not intervene with radiation, the net rate of radiation heat transfer between these two surfaces is given by:

(2.5)

CHAPTER 2. Electromagnetic and thermal dosimetry at millimeter waves: effects of thermal convection during in vitro exposure of cells

Figure 2.2: Radiation exchange between a surface and large surroundings.

2.2.3 Convection Convection is the mechanism of heat transfer between a surface and the adjacent gas or liquid in motion. The energy is transferred by the combination of two processes including i) the heat conduction due to the random molecular motion, and ii) the bulk, or macroscopic motion of the fluid. This fluid motion is associated with the fact that, at any instant, large numbers of molecules are moving collectively or as aggregates. Such a motion, in the presence of a temperature gradient, contributes to heat transfer. Convection heat transfer between a fluid in motion and a bounding surface at different temperatures is represented in Figure 2.3, where a hot solid surface is in contact with a fluid (e.g., air). In the Figure 2.3 two regions, defined respectively by the velocity and the temperature of the air, are shown. They represent the hydrodynamic or velocity boundary layer and the thermal boundary layer, respectively. At the interface between the two media, where the velocity of the fluid is zero, heat from the hot solid surface is transferred to the closer layer of the fluid only by heat conduction due to the random motion of the molecules. The contribution due to the liquid bulk motion originates from the fact that the velocity boundary layer grows as a function of the x direction from zero to a finite value . The macroscopic motion of the air enhances heat transfer considerably by removing the heated air near the surface and replaces it by the cooler fluid further away. In particular, the natural buoyancy forces due to the density gradients in the fluid, induced by temperature changes, determine the rise of the warmer air (and thus lighter) near the surface and the fall of the cooler (and thus heavier) air to fill its place. The faster the fluid motion, the greater the convection heat transfer.

CHAPTER 2. Electromagnetic and thermal dosimetry at millimeter waves: effects of thermal convection during in vitro exposure of cells

Convection heat transfer is expressed by Newton's law of cooling as:

(2.6)

where (W) is the heat flux, (W/m²·°C) is the convection heat transfer coefficient, A (m) is the surface area through which convection heat transfer takes place, (K) is the temperature of the surface, and is the fluid temperature in a point sufficiently far from the hot surface. The presence of bulk motion of the fluid if enhances the heat transfer between the solid surface and the fluid, it also complicates the determination of heat transfer rate. The coefficient is not a property of the fluid, but it is an experimentally determined parameter whose value depends on all the variables influencing convection such as the surface geometry, the nature of fluid motion, the properties of the fluid, and the bulk fluid velocity.

Figure 2.3: Convection heat transfer from a hot surface to air.

Convection is classified as natural (or free) and forced depending on how the fluid motion is initiated. In natural convection, the fluid motion is caused by the natural buoyancy forces induced by density gradients in the fluid due to temperature differences. In forced convection, the fluid is forced to move by an external mean such as a fan, a pomp, or the wind; the use of fun or in general blowers enhances by order of magnitude the convection. Typically the greater the fluid flow, the stronger the convection. Natural convection has been widely investigated and well discussed in literature [22]–[28].

CHAPTER 2. Electromagnetic and thermal dosimetry at millimeter waves: effects of thermal convection during in vitro exposure of cells

Since the solution of the differential equations describing convective heat transfer is very complicated, the dimensionless groups such as Rayleigh (Ra), Grashof (Gr), Prandtl (Pr), and Nusselt (Nu) are used to facilitate the characterization of the natural convective flow [26]. They are obtained by non-dimensionalizing the governing equations of the heat transfer mode and combining the variables together in order to reduce the number of total variables. The definition of such numbers allows the estimation of the type of fluid flow as well as the heat transfer coefficient , used for the quantification of the convection heat transfer rate. Fluid flows are generally classified into laminar, when the fluid motion is highly ordered and it is characterized by smooth streamlines, or turbulent when the fluid motion is highly disordered and associated to high velocities (Fig. 2.4).

Figure 2.4: Isotherms in natural convection in laminar flow (left) and turbolent flow (right).

The threshold separating the quiet from convective regimes is defined by the Ra number:

(2.7)

where g (m/s2) is the gravitational acceleration, β (1/K) is the coefficient of volume expansion, x (m) is the characteristic length of the geometry, ΔT (°C) is a driving temperature difference between the temperature of the surface and fluid sufficiently far from the surface, ν (m²/s) is kinematic viscosity of the fluid, and α (m²/s) is the thermal diffusivity. The latter is a measure of the ability of a material to conduct thermal energy relative to its ability to store it (i.e., the rate of transfer of heat

CHAPTER 2. Electromagnetic and thermal dosimetry at millimeter waves: effects of thermal convection during in vitro exposure of cells of a material from the hot end to the cold end). It is equal to k/ρ∙cp, i.e., the thermal conductivity k

(W/(m∙°C)) divided by density ρ (kg/ ) and specific heat capacity of liquid cp (J/(kg∙°C)) [22]. Materials with large α will respond quickly to changes in their thermal environment, while materials with small α will respond more sluggishly, taking longer to reach a new equilibrium condition. For Ra numbers below some critical value, the fluid is stable. When Ra number is above critical value the onset of natural convection occurs. For convection of fluid between solid boundaries the critical Ra number was calculated to be 1708.8 [29]. For other combinations of boundaries the critical Ra number is < 1708.8 [30]. The Gr number is defined by the ratio of the buoyant forces on the fluid to the viscous forces:

(2.8)

where the parameters are the same as for Equation (2.7). Gr number is used to define the convection type, i.e., laminar or turbulent. For Gr number exceeding 109 the convection flow is considered to be turbulent [25]. The higher the Gr number, the higher the fluid movement, and therefore the convection. The coefficient of volume expansion, β, is a measure of the amount of the density (or volume) changes of a substance in response to a change in temperature at constant pressure. It assumes particular importance in the study of the free convection as the latter is triggered by the temperature variations, resulting in density variation of the fluid. It is expressed as:

(2.9)

where in addition to parameters already defined in Equations (2.7) and (2.8), (kg/ ) is the density and (K) is the temperature of the quiescent fluid away from the surface. A large value of β in the fluid means a large change in density. The larger the temperature difference between the fluid adjacent to a hot (or cold) surface and the fluid away from it, the larger the buoyancy force and the stronger the natural convection currents, and thus the higher the heat transfer rate. The Pr number is a measure of the ratio between the molecular diffusivity of momentum ν, and the molecular diffusivity of heat α:

(2.10)

CHAPTER 2. Electromagnetic and thermal dosimetry at millimeter waves: effects of thermal convection during in vitro exposure of cells where the parameters are the same as for equations (2.7) and (2.8), and µ (cP) is the dynamic viscosity. It is worth to note that whereas the Ra and Gr numbers are expressed as a function of the geometrical dimension of the system (i.e., x), the Pr number does not contain such length scale and it is dependent only on the fluid and the fluid state. The Pr number provides a measure of the relative effectiveness of momentum and energy transport by diffusion in the velocity and thermal boundary layers, respectively. The Pr numbers of gases are about 1, indicating that both momentum and heat dissipate through the fluid at about the same rate. Heat diffuses very quickly in liquid metals (Pr << 1) and very slowly in oils (Pr >> 1) relative to momentum. The Pr numbers of fluids range from less than 0.03 for liquid metals to more than 10∙104 for heavy oils, while it is in the order of 10 for water [13]. The Nu number is a measure of the ratio of convection to pure conduction heat transfer under the same conditions:

(2.11)

where is the convective heat transfer coefficient, k the thermal conductivity, and x the characteristic length. The Nu number represents the enhancement of heat transfer through a fluid layer as a result of convection relative to conduction across the same fluid layer [27]. The larger the Nusselt number, the more effective the convection. A Nu number equal to 1 for a fluid layer represents the heat transfer across the layer by pure conduction. Due to the difficulties associated to the determination of through the solution of the governing equations, a simple empirical correlation is typically used for the determination of Nu number as a function of the Gr, Pr, and Ra numbers:

(2.12)

The values of the constants C and n depend on the geometry of the surface and the flow regime, characterized by the range of the Rayleigh number. The value of n is usually 1/4 for laminar flow and 1/3 for turbulent flow, while the value of the constant C is normally less than 1 [27].

2.3 Experimental set-up

CHAPTER 2. Electromagnetic and thermal dosimetry at millimeter waves: effects of thermal convection during in vitro exposure of cells

The exposure system used to carry out our experiments consists of three main units: 1) an exposure chamber, 2) a signal generator sub-unit, and 3) a temperature interface. Schematic representation of the exposure system is given in Figure 2.5.

Figure 2.5: Outline of the exposure system: exposure chamber (incubator), signal generation unit, and temperature recording instrumentation.

Exposure chamber. All exposures were carried out inside a MEMMERT UNE 400 incubator (Memmert GmbH, Schwabach, DE). Real view inside the incubator in the exposure configuration is given in Figure 2.6. The temperature of the incubator was set at 32°C to compensate for the rapid temperature rise during the first minutes of exposure. Samples under test (SUTs), described in the following section, were placed into one well of a 12-well tissue culture plate (TCP) (353072, Microtest 96, Becton Dickinson, Franklin Lakes, NJ) made of polystyrene. Each well was 22.09 mm in diameter. Due to the shallow penetration of the MMW, exposures were performed from the bottom by an open-ended rectangular waveguide (WG) WR15 (aperture size 3.81 × 1.905 mm²) located 5 mm from the TCP. The latter was set on a plastic support with a 3.5 cm hole in diameter (Fig. 2.6b), centred with the exposed well. Considering that small shifts (of the order of mm) from the center of the exposure result in large variations of SAR and temperature distributions, to avoid movements, TCPs were set into a plastic frame glued on the plastic support (Fig. 2.6c). Due to the very local exposure and high free-space losses MMW, the contributions of the multi-path reflections inside the incubator is extremely small, therefore MMW absorber is only at the bottom of the incubator.

CHAPTER 2. Electromagnetic and thermal dosimetry at millimeter waves: effects of thermal convection during in vitro exposure of cells

(a)

(b) (c)

Figure 2.6: (a) Real view of the exposure chamber (inside). (b) Plastic TCP support with a 3.5 mm hole diameter aligned with the open ended rectangular WG. (c) Zoom on the TCP embedded in the plastic frame attached on the TCP support.

Signal generator unit. High power generator (QuinStar Technology, Torrance, CA) operating at 58.4 GHz was used as a source. In the PW regime, a programmable power supply HMP (Hameg Instruments, Hampshire, UK) supplying control voltage was used for amplitude modulation of

CHAPTER 2. Electromagnetic and thermal dosimetry at millimeter waves: effects of thermal convection during in vitro exposure of cells

MMW to create thermal pulses with duration ranging from 1.0 s to 6.5 s. The input power of the open-ended WG was measured before experiments using a V-band Agilent V8486A power meter.

Figure 2.7: MMW high power signal source subsystem (left) operating at 58.4 GHz equipped with power supply subsystem packaged in a separate enclosure and connected using a cable/wire assembly (top right). Programmable power supply (bottom right) used for the amplitude modulation of the MMW to generate thermal pulses with different durations.

Temperature interface. Temperature measurements were performed by using a micro- thermocouple (TC) connected to a portable computer for the real-time temperature acquisition.

2.4 Samples under test

Development of convection was analyzed in four models including distilled water, cell culture medium, monolayer of melanoma cells covered by culture medium, and agar gel. The volume of the SUTs ranged from 0.5 to 4 ml corresponding to 1.3 and 10.4 mm height, respectively.  SUT 1. Distilled water, provided by the microelectronic department of the IETR, with a resistivity of 15 MΩ. At MMW, the electromagnetic properties of the culture medium are close to those of pure water as discussed more in detail hereafter (section 1.8.1, Heating in different media).  SUT 2. Culture medium. Dulbecco’s modified eagle medium (DMEM) made up by amino acids, vitamins, inorganic salts, and other components (D-glucose, phenol red, sodium pyruvate) supplemented with 8% of fetal calf serum, 1% antibiotics, and 1% L-glutamine.  SUT 3. Melanoma cell monolayer attached to the bottom of the exposed well and covered by the culture medium. A375 melanoma cells [31] were seeded at concentration of cells in 2 ml of the same culture medium as SUT 2. Before exposure the medium was replaced by DMEM 94

CHAPTER 2. Electromagnetic and thermal dosimetry at millimeter waves: effects of thermal convection during in vitro exposure of cells

without sodium bicarbonate containing 4.6 mM of Hepes (Thermo Fisher Scientific, Waltham, MA, USA) to maintain constant pH in the non-gassed incubator of the exposure system [32].  SUT 4. Agar gel (98.5% of distilled water and 1.5% of agar). It was used as a reference water-equivalent model without convection. Except viscosity, both electromagnetic [33] and thermophysical parameters [34] of the agar phantom closely match those of pure water. The fabrication procedure of the agar phantom is outlined hereafter. Distilled water is heated up to boiling. Then agar is progressively added (Fig 2.8a) and stirred until it dissolves into water. The obtained mixture was poured with a pipette into a well of a 12-well TCP where a thin TC was previously attached to the bottom (Fig 2.8b). Afterwards, the sample was left at room temperature to solidify.

(a) (b)

(c) Figure 2.8: (a) Mixing of the agar powder in the boiling water. (b) Agar gel poured into one well with the TC attached in the center at the bottom of the plastic well. (c) Agar phantom at different heights (i.e., 1.5 ml and 5 ml, left and right respectively).

All the SUTs were incubated for 1 h before the exposure to allow the temperature equilibrium at 32 °C.

CHAPTER 2. Electromagnetic and thermal dosimetry at millimeter waves: effects of thermal convection during in vitro exposure of cells

2.4.1 Electromagnetic properties The complex permittivity of water, culture medium, and agar gel was measured at room temperature using an open-ended coaxial probe DAK-1.2E (SPEAG, Zurich, CH) with the procedure described in Appendix B. The electromagnetic properties of the SUTs are given in Table 2.1 at 58.4 GHz at 22 °C due to the fact that permittivity measurements at temperatures exceeding the room temperature result in increased uncertainty. Differences among the models are within the 10%, 4%, and 8% for the real part of the complex permittivity (i.e., ε), the electrical conductivity (i.e., σ), and loss tangent (i.e., tanδ) respectively. Note that, according to the free water temperature- dependent model proposed by Ellison [35], variation of electromagnetic properties in the 22–52°C range (room temperature and maximum temperature reached during long exposure of the SUTs, respectively) results in SAR variation of less than 2%. Complex permittivity of polystyrene was determined by using a free-space technique with a transmission/reflection quasi-optical setup and ABmm millimeter-wave vector network analyzer [36].

Table 2.1. Relative permittivity ε, conductivity σ, and loss tangent tanδ of considered materials

ε σ (S/m) tanδ

Water 12.67 66.79 1.62 Culture 11.25 64.16 1.75 medium Agar gel 12.29 64.94 1.625 Polystyrene 2.56 0.008 9.76∙10-4

2.4.2 Thermo-physical properties Thermal (i.e., heat capacity and thermal conductivity) and mechanical (i.e., viscosity) properties of water, culture medium, and agar gel are provided in Table 2.2. The firsts were measured in the Laboratoire de thermique et d'energie de Nantes (Appendix C), while values of viscosity were taken from literature because of the poor precision obtained by using the instrumentation provided by the

Institut of Physique of Rennes (IPR) as described in Appendix D. Specific heat capacity and thermal conductivity k of water and culture medium were measured with a differential scanning calorimeter (DSC Q200, TA Instruments, New Castle, DE) and a thermal conductivity analyzer (C- Therm Technologies, New Brunswick, CDN), respectively, within the 32–52 °C range (minimum and maximum temperatures in our experiments). These instrumentations allowed us to perform 96

CHAPTER 2. Electromagnetic and thermal dosimetry at millimeter waves: effects of thermal convection during in vitro exposure of cells measurements even at higher temperature than room temperature without uncertainties. In the 32– 52 °C temperature interval, the specific heat capacity variation is less than 1%, and the thermal conductivity increases by 2%. Heat capacity and thermal conductivity values of the agar gel are well known. They are close to those of water as shown in several studies [37]–[40]. Density  of water, culture medium, and agar gel are taken from [4], [13], and [34], respectively. Viscosity η (cP) of the liquid samples plays an important role in convection initiation, which occurs as a result of the motion of a fluid due to the density changes arising from heating [39]. The viscosity of water has been characterized in several studies [13], [40]. The viscosity of the culture medium is 7% higher compared to that of water [41]. The viscosity of the agar gel is much higher compared to liquids. In the 32–52 °C interval, the agar gel can be considered as solid and its viscosity diverges towards infinite [42].

Table 2.2. Thermal and mechanical properties of water, culture medium, and agar gel at 32 °C

k ρ η

(J/(kg∙°C)) (W/(m∙°C)) (kg/ ) (cP) Water 4168.9 0.625 1000 0.69

Culture medium 4067.3 0.619 1000 0.75

Agar gel 4000 0.6 1000

2.5 Temperature measurements

Local temperature measurements during in vitro experiments at MMW is very challenging due to the fact that electromagnetic power is locally concentrated within few tenth of mm close to the surface of the exposed sample inducing high temperature gradients. As detailed in the Chapter 1 (paragraph 1.2.5, Techniques for measurement of MMW-induced heating), conventional dosimetric MMW techniques (i.e., infrared thermometry [33], fiber optic thermometry [43], and thermosensitive liquid crystals [44]) fail to adequately measure the temperature rise induced in the SUTs during in vitro exposure. On the contrary, microscale TC allows for accurate local measurement of temperature with a time constant of the order of µs. This technique is particularly well adapted to measure the sharp MMW-induced thermal pulses [7].

CHAPTER 2. Electromagnetic and thermal dosimetry at millimeter waves: effects of thermal convection during in vitro exposure of cells

2.5.1 Operation principle of thermocouple A TC is a temperature-voltage transducer, which is made by welding two different metals at one end (Fig. 2.9a). The joined end is referred to as the hot junction and it is used to measure the temperature of interest, while the other is referred to cold end or cold junction, and it is placed at a known reference temperature. The working principle of the instrument is based on the combined Seedbeck and Peltier effects (Fig. 2.9b,c) which are different manifestations of the same physical process. The Seedback effect is the direct conversion of thermal to electrical energy: when two wires of dissimilar metals (i.e., alumel and chromel) are joined at both ends, in a closed loop configuration, and one of this ends is heated, a continuous current, I, flows in the thermoelectric circuit which is proportional to the temperature difference between the two junctions. The Peltier effect is the direct conversion from electrical to thermal energy: when two different metals are welded at both the ends and a voltage is applied breaking the loop configuration, the electrical current flowing through the junction connecting the two materials will emit or absorb heat per unit time at the junction, to balance the difference in the chemical potential of the two materials. If the voltage is small, the heat transferred is linearly proportional to the voltage (i.e., ΔV). Fig. 2.9a shows the TC in the measurement configuration: when a temperature difference (Tc) is recorded at the two junctions, a voltage will develop (Seedback effect) depending on it. If the temperature difference is small, the voltage is nearly linearly proportional to the temperature difference (Peltier effect) [45].

(a)

CHAPTER 2. Electromagnetic and thermal dosimetry at millimeter waves: effects of thermal convection during in vitro exposure of cells

(b) (c) Figure 2.9: (a) K-type thermocouple (chromel-alumel) in the standard measurement configuration. The measured voltage can be used to calculate the temperature Tc, provided that temperature Tref is known. (b) Schematic of TC thermal to electrical Seebeck effect. (c) Schematic of TC electrical to thermal Peltier effect.

2.5.2 Impact of thermocouple on temperature measurements The presence of metallic structures (such as TC) in the electromagnetic field may distort the SAR and consequently the temperature in the exposed sample. However, the possibility to use small TC, with a lead diameter < 1 mm, for measurements of the temperature rise during MMW exposure has been evaluated by previous studies performed by Alekseev and Ziskin [46], [47], and confirmed experimentally in this work. The interaction between a TC and the electromagnetic field may determine numerous artefacts including 1) stray pickup in the input circuits of the measuring equipment, 2) current induction in thermoelectrode loops, and 3) dipole current induction in the TC depending on its orientation to the E-field. The same authors [46] demonstrated that 1) the input circuit interference is easily abolished by careful shielding, 2) twisted wires of the thermoelectrodes prevent induced currents, and 3) TC orientation perpendicular to the E-field removes dipole currents. The correct TC positioning with respect to the E-field is of outmost importance for the artefact-free measurements of the MMW-induced heating in the exposed object. The SAR distributions were calculated for two different configurations, including the perpendicular and parallel orientations of the TC to the E-field inserted either in water [46] or skin [47]. Specifically, in [47], numerical calculations were performed in a model of skin in the presence of a TC modeled as a metallic cylindrical structure of 0.1 mm diameter and length of 2 mm. Skin was modeled as a rectangular block of homogeneous medium of 1 mm thickness with the electromagnetic properties of the skin, covering an area of 24 mm2 (Fig. 2.10a,c). The skin-TC specimen was exposed to a normally incident plane wave of 42 GHz at IPD = 10 mW/cm2. Results demonstrated that the insertion of TC in skin produced a strong localized disturbance of otherwise uniform SAR distribution in the exposed sample, both for the parallel and perpendicular orientations of the TC in respect to the E-field. Results showed that inside the TC, SAR values were close to zero because of poor metal penetrability for MMW, while outside the TC the SAR was localized in the immediate 99

CHAPTER 2. Electromagnetic and thermal dosimetry at millimeter waves: effects of thermal convection during in vitro exposure of cells proximity of the TC tip along the E-plane (Fig. 2.10b,d). However, significant differences were observed among the parallel and perpendicular orientations. In particular, when the TC was oriented parallel to the E-field (Fig. 2.10a,b), SAR was highly distorted around the TC tip by about 24-fold (Fig. 2.10b) compared to the control where no TC was inserted. On the contrary, when the TC was set perpendicular to the E-field (Fig. 10c,d), the SAR in proximity of the TC tip was only weakly distorted, with differences within the 1% compared to the control (dotted line of Fig. 2.10d).

(a) (b)

(c) (d) Figure 2.10: SAR distribution in skin exposed to 42 GHz at IPD = 10 mW/cm2 with a TC oriented parallel (a,b) and perpendicular (c,d) to the E-field. (a) Grayscale plot on the surface of the skin. (b) SAR profiles along lines 1 and 2. (c) Grayscale plots on the surface and through skin depth along line 1. (d) SAR profiles along lines 1, 2 on the surface, and 3, 4 in depth. Dotted line represents the control profile along line 1 without the TC [47]. 100

CHAPTER 2. Electromagnetic and thermal dosimetry at millimeter waves: effects of thermal convection during in vitro exposure of cells

The same authors, in another study [46] computed SAR distributions in water in the presence of a TC by means of analytical calculations. The TC was modeled as a metallic cylindrical structure of 0.1 mm diameter in a 2.6 × 2.6 mm sample simulating water (insert to Fig. 2.11a). The objects were exposed to a plane wave at 42.25 GHz, at IPD = 13.2 mW/cm2. The TC was oriented perpendicular to the E-field. Under these conditions, the SAR distributions were calculated in the directions parallel and perpendicular to the E-field (Fig. 2.11). Results were similar to those reported in [47] confirming that a thin (diameter < 0.1 mm) metallic TC highly distorts the SAR distribution around the TC tip, along the E-plane. In particular, in proximity of the TC tip along the E-field (E∥ in Fig. 2.11) the maximum SAR values exceeded the average uniform SAR in the absence of the TC by more than four times, while notably reduced values, close to zero, were found in the opposite direction (E⊥ of Fig. 2.11). These results were similar to those obtained in this study by means of numerical simulations (2.6, Electromagnetic dosimetry). Figure 2.11c shows that the size of the distortion area depends on the TC diameter, i.e., the smaller the diameter, the smaller the distorted area.

(a)

101

CHAPTER 2. Electromagnetic and thermal dosimetry at millimeter waves: effects of thermal convection during in vitro exposure of cells

(b) (c) Figure 2.11: Normalized SAR distribution in the area of the TC located in the center of the water sample as shown in insert to (a). TC was modeled as a cylindrical object exposed to MMW at 42.25 GHz at IPD = 13.2 mW/cm2. (a) Three-dimensional SAR plot. E indicates the direction of the E-field. (b) Profiles of SAR distributions. E∥ and E⊥ indicate the directions parallel and perpendicular to the E-field, respectively. (c) Normalized SAR distribution in water in the vicinity of 0.05, 0.1, and 0.2 mm diameter TCs for E∥ and E⊥ [46].

There is no absorption of MMW energy inside the TC, thus the temperature rise measured by the TC during exposure is due to the heating of the solution in contact with the TC. Therefore, the knowledge of the average absorbed MMW energy in the sample around the TC may provide qualitative information about the possible impact of the presence of the TC on the heating of the sample. In their works [46], [47], the authors found that in the case of parallel orientation of the TC, the average SAR in proximity of the TC was significantly higher than SAR in the control. This enhanced absorption of the MMW energy in the parallel configuration may lead to overheating, especially at the TC tip. On the contrary, for the perpendicular configuration, authors found negligible differences between the SAR averaged over the area surrounding the TC and the SAR in the control without TC, with differences less than 1%. This suggests that the induced-heating in the exposed sample should be the same with and without the TC. This hypothesis was experimentally verified in another study by Alekseev and Ziskin [48], measuring the temperature elevation both by means of a thin TC (0.1 mm diameter) and infrared (IR) camera. In this case, water-based phantoms were exposed to 42.25 GHz at output powers ranging from 20 mW to 50 mW. It was demonstrated that the ΔT profiles in the exposed phantoms with and without the TC were identical except in the region of the TC itself (Fig. 2.12a), where the recorded ΔT was lower. However, the heating kinetics measured with the IR camera and TC showed negligible differences in the induced heating, within the TC accuracy, i.e., 0.1 °C (Fig. 2.12b).

102

CHAPTER 2. Electromagnetic and thermal dosimetry at millimeter waves: effects of thermal convection during in vitro exposure of cells

To summarize, results of these studies showed that TCs with diameter ≤ 0.1 mm represent optimal devices to measure the temperature elevation during MMW exposure providing that the probe is inserted perpendicular to the E-field. On the contrary, induction of dipole currents in the TC and severe distortion of the SAR distribution make the parallel orientation of TC to the E-field unsuitable for temperature measurements under MMW exposure.

(a) (b) Figure 2.12: (a) ΔT profiles of the IR image obtained in the presence (dots) and absence (solid line) of a TC, respectively in water-based phantoms exposed at 42.25 GHz. (b) Heating kinetics measured with a TC (solid line) and IR-camera (dots) in the place of a TC tip [48].

In our experiments, SUTs were exposed to MMW at 58.4 GHz. Note that at this frequency, the length of a standard dipole antenna in water is around 0.8 mm (λwater/2 [49]). Therefore, TC with dimension < 0.1 mm (i.e., 75 µm in this study) cannot act as a resonant dipole. Because of poor metal penetrability of MMW, the temperature elevation recorded during the exposure is mainly due to the heating of the exposed SUTs in contact with the TC. Therefore, the presence of dipole currents can be easily verified experimentally by performing measurements in free space by exposing only the TC without the TCP. We tested the susceptibility of TC to induced currents and electromagnetic interference with our exposure system, both under the parallel and perpendicular configurations. The TC leads were attached to the plastic support located inside the incubator, aligned to the axis of the open-ended WG (Fig. 2.13) and exposed to PW with the pulse duration of 1.5 s at the peak generator power of 4.2 W. Under the parallel configuration the temperature measured by the TC (75 µm diameter leads) was high due to possibly induced currents which may lead to the overheating of the TC tip (Fig. 2.14a). In this case, TC reading is about 20-fold higher compared to the temperature recorded by the TC with the perpendicular orientation of its leads. The measurements showed that the perpendicular orientation of the TC leads with respect to the E-field 103

CHAPTER 2. Electromagnetic and thermal dosimetry at millimeter waves: effects of thermal convection during in vitro exposure of cells reduced the induced currents, as only negligible temperature increase is measured by the TC (Fig. 2.14b). Maximum deviation of about 5% was observed compared to measurements in absence of exposure (Fig. 2.14c).

(a) (b)

Figure 2.13: TC in free space configuration oriented (a) parallel and (b) perpendicular to the E-field.

(a) (b)

104

CHAPTER 2. Electromagnetic and thermal dosimetry at millimeter waves: effects of thermal convection during in vitro exposure of cells

(c) Figure 2.14: Temperature dynamics in free space in PW regime with a pulse duration of 1.5 s at 4.2 W for (a) the parallel and (b) perpendicular orientations of the TC leads. (c)Temperature dynamics for the perpendicular orientation of the TC leads during exposure to PW with a pulse duration of 1.5 s at 4.2 W (black line), and in absence of the exposure (red line).

2.5.3 Measurements of thermal pulses The possibility to measure short thermal pulses induced by amplitude-modulated MMW was analyzed in detail in [7] by comparing temperature dynamics recorded with FO thermometer with relatively big diameter of the probe (i.e., 1.15 mm), and TCs with two different sizes of the leads (i.e., 25 µm and 75 µm) as shown in Figure 2.15. In this case, thermal pulses of 1 s at peak generator power of 4.2 W and pulse repetition rate at 0.05 pps were generated using the same technique as the one described in paragraph 2.3 (section Signal generator unit). The analysis of the temperature dynamics locally recorded during PW regime as a function of the diameter of the thermosensor revealed that the smaller the TC, the more accurate is the temperature measurement. The study showed that both the shape and the amplitude of the thermal pulses were strongly dependent on the method of temperature measurement. Sharp and high amplitude thermal pulses were measured by both the TCs. The smallest TC (i.e., 25 µm) allowed to measure the maximum amplitude of the peak of 10 °C. Temperature amplitude recorded with TC of bigger diameter (75 µm) was about 20% smaller (8 °C) than the one recorded with TC of smaller diameter. Temperature pulses obtained with FO probe had much lower amplitudes and longer initial phases of reaching the maximal amplitude. Amplitude of thermal pulses were approximately 8 and 6 times lower than those measured with 25 µm and 75 µm leads diameters, respectively. The failure of FO devices to appropriately measure rapid thermal pulses may be possibly due to the to the great time constant

105

CHAPTER 2. Electromagnetic and thermal dosimetry at millimeter waves: effects of thermal convection during in vitro exposure of cells

(0.6 s) of their input which may possibly play the role of a low-pass filter reducing amplitude of pulses and shifting their position on the time axis.

(a) (b) Figure 2.15: Temperature pulses measured at bottom of a well with FO sensor, 75 µm, and 25 µm TC. Liquid volume is (a) 1.5 ml and (b) 3 ml [7].

106

CHAPTER 2. Electromagnetic and thermal dosimetry at millimeter waves: effects of thermal convection during in vitro exposure of cells

2.5.4 Thermocouple configuration In this study, local temperature at the bottom of the well was measured by using a K type TC (RS Components, Corby, UK) probe with the lead diameter of 75 µm. The lead was carefully attached to the bottom of the well with a tape, leaving the TC tip free. To record the temperature we used Thermocouple Reference design (Microchip Technology, Chandler, AZ) with sampling interval of 156 ms. The software Thermal Menagment Utility (Microchip Technology) was used to visualize and record real-time temperature elevation during the exposure.

(a) (b) Figure 2.16: Temperature interface. (a) K-type TC with a lead diameter of 75 µm, connector (yellow), and microchip (green). (b) Real-time temperature measurement with Thermal Menagment Utility during PW exposure.

The tip of the TC was aligned with the exposure beam axis, which coincided with the axis of the exposed well. The leads of the TC laid on the bottom of the well, perpendicular to the E-plane (Fig. 2.17) to prevent possible artefacts in temperature measurements related to induced currents [46]– [48], [50]–[53].

107

CHAPTER 2. Electromagnetic and thermal dosimetry at millimeter waves: effects of thermal convection during in vitro exposure of cells

(a) (b) Figure 2.17: TC setting. (a) Real view from the top of TC aligned with the exposure beam axis. (b) Schematic of the top view.

2.5.5 Reproducibility Reproducibility of the temperature measurements represented a critical point of this work. The reliability of the results presented in this Chapter, as well as the correct interpretation of the results shown in Chapter 3, largely depend on the reproducibility of the temperature dynamics recorded during the exposure of the SUTs. As already mentioned, at MMW even small shift of few mm between the antenna and the sample may result in large variation of the SAR and temperature distributions. Therefore, to guarantee that all the SUTs were exposed under identical exposure conditions, the correct positioning of the sample, with respect to the center of the exposure, corresponding to the maximum SAR and induced-temperature rise, was fundamental. Moreover, as already shown (section 2.5.2, Impact of thermocouple on temperature measurements) the correct positioning of the TC with respect to the E-field is of outmost importance for the artefact-free measurements of the MMW-induced temperature elevation. For this reason, before each experiment both TC position and orientation were verified. First, each TC was calibrated and its leads were joined. For measurements, TC leads were attached to the bottom of the well. The correct attachment was verified at the beginning of each exposure, as small detachment may determine the acquisition of temperature rise in points different from the center of the exposure beam. In each set of experiments, TC was aligned with the axis of the open-ended WG with its leads perpendicular to the E-field (Figure 2.17). Preliminary measurements were performed to assure the alignment of the TC (and the TCP) with respect to the WG axis following the procedure described hereafter. As the maximum peak temperature induced by a 1.5 s pulse at 4.2 W was 10 °C, the plate was considered aligned when the recorded

108

CHAPTER 2. Electromagnetic and thermal dosimetry at millimeter waves: effects of thermal convection during in vitro exposure of cells temperature was equal to 10°C. The presence of a plastic frame (Fig. 2.6c) allowed to fix the TCP in the desired position assuring its stability during the whole set of measurements. Under this condition, the TCP support was carefully maintained in the same location during all the experiments in the same set of exposures. After each measurement TCP was removed from the incubator without moving its support, and liquid (or agar gel) was removed too. The exposed well containing the TC was rinsed with alcohol and dried before the successive measurement. The change of the sample after each exposure was fundamental to assure that the desired liquid volume was exposed (i.e., volume may decrease due to evaporation). Results obtained by using the above mentioned protocol were compared to those obtained by re-setting and re-aligning TC, TCP, and TCP support for each experiment (i.e., the plastic frame was not used to fix the TCP in the desired position). Negligible differences, within the standard deviation of the measures, were found confirming the reliability of the methodology employing the plastic frame to set the sample. In the case of the PW exposure, 1 mm shift from the center resulted in a decrease of the peak amplitude of the thermal pulse by about 1 °C (i.e., about 10%). At only 3 mm from the center of the exposure, peak amplitude of the heat pulse decreases by about 57% with respect to the peak values on the axis, both along E- and H-planes (Fig. 2.18). This means that small shift of the TCP with respect to the center may introduce great errors in the recording of the temperature dynamics, impacting the interpretation of the results when comparing measurements for the same distance from the center. In the case of CW exposure, shift from the center was less critical as the temperature was mostly homogeneous up to about 2.5 mm from the center (see Fig. 3.14, in section 3.3.1, 1.5 s heat pulses).

109

CHAPTER 2. Electromagnetic and thermal dosimetry at millimeter waves: effects of thermal convection during in vitro exposure of cells

Figure 2.18: Normalized peak amplitude of thermal pulse with 1.5 s duration at 4.2 W as a function of the distance from the center of the exposure (i.e., d= 0 mm), along the E- (gray line) and H-planes (black line).

Preliminary experiments showed that the reproducibility was largely influenced also by the initial temperature of the sample before exposure. They showed that when the sample was exposed at different temperature from the one of the incubator (32°C), the temperature dynamics of the SUTs was either faster or lower, with differences within the 10–15%. Therefore, all the SUTs were exposed only once they reached the temperature stability after about 1h of incubation at 32 °C. In this study, reproducibility was assured by repeating measurements at least 4 times per each condition if not stated otherwise. Temperature differences observed among the SUTs exposed under the same experimental conditions were within the accuracy of the measurements (i.e., within 2%). An example of the reproducibility (for the sake of brevity only two experiments/conditions are shown) of temperature dynamics for CW and PW regimes is represented in Figure 2.19 for water, agar gel, and culture medium obtained for 6 independent measurements.

110

CHAPTER 2. Electromagnetic and thermal dosimetry at millimeter waves: effects of thermal convection during in vitro exposure of cells

(a) PW (Water)

(b) CW (Agar gel)

(c) CW (Culture medium) Figure 2.19: Reproducibility of the temperature dynamics under PW and CW exposures in (a) water during PW exposure, (b) agar gel, and (c) culture medium during CW exposure.

111

CHAPTER 2. Electromagnetic and thermal dosimetry at millimeter waves: effects of thermal convection during in vitro exposure of cells

2.6 Electromagnetic dosimetry

Numerical simulations were performed to analyze both the SAR and temperature distributions inside the exposed well, as well as to further understand the impact of thin TC (leads diameter < 0.1 mm) on the temperature measurements. The finite integration technique (FIT) solver of CST Microwave Studio 2018 (CST-Computer Simulation Technology, Dassault Systemes, DE) was used for the electromagnetic analysis. The exposure scenario (Fig. 2.20) is represented by one well of the TCP and the feed antenna to reduce the computational volume represented in each simulation by about 45 million mesh cells. This assumption is justified by the fact that in MMW exposures the absorbed energy is mainly localized at the bottom of the plate close to the well axis, making negligible the contribution of reflections from the neighboring empty wells. In order to carefully account for high SAR gradients at the bottom of the exposed liquid, non-uniform adaptive meshing was used. The simulations were performed using a mesh with a cell size ranging from 2 µm in liquid up to 340 µm in free space. Perfectly matched layers absorbing boundary conditions were used. Minimal distance between the culture plate and boundaries was set to λ0/4. In simulations, the permittivity and conductivity (Table 2.1) of the SUTs were considered to be temperature independent taking into account the small variation of SAR (i.e., 2%) in the considered 22–52 °C temperature interval.

112

CHAPTER 2. Electromagnetic and thermal dosimetry at millimeter waves: effects of thermal convection during in vitro exposure of cells

Figure 2.20: CAD model. Zero of the coordinate system is located at the center of the lower surface of the well bottom. Zoom of the TC is shown on the left.

Table 2.3 summarizes the peak values of SAR computed at 58.4 GHz for WG input power of 195 mW (the same power level as the one used to experimentally determine SAR based on temperature measurements [section 2.7, Temperature-based retrieval of SAR]). From the electromagnetic point of view the differences between the four SUTs are almost negligible (maximum deviation of SAR with respect to that in water is within 2% for the culture medium with the cell monolayer).

Table 2.3. Computed (without TC) and measured peak values of SAR at 58.4 GHz

Deviation in SAR computed SAR measured Corrected SAR peak respect to SAR in peak peak (W/kg) (W/kg) (W/kg) water Water 7767 0% 6484.1±380.4 7780.9±456.4 Culture medium 7832 1% 6583.2±522.6 7899.8±627 (without cells) Culture medium 7910 2% 6686.8±701.0 8024.1±841.2 (with cells) Agar gel 7725 0.5% 6427.0±163.2 7712.4± 195.8

Figure 2.21 shows the SAR distribution in water at the interface with the bottom of the well at z=0.02 mm, corresponding to the maximum SAR along E- and H- planes, and on its axis. SAR

113

CHAPTER 2. Electromagnetic and thermal dosimetry at millimeter waves: effects of thermal convection during in vitro exposure of cells rapidly decreases along all the directions; for instance, the SAR drop from 50% to 10% levels corresponds to (x = 3.8, y = 2.5, z = 0.1 mm) and (x = 8.7, y = 5.3, z = 0.37 mm), respectively.

(a)

(b) Figure 2.21: Computed SAR without TC (in black): (a) at the interface with the bottom of the well in E and H planes at z = 0.02 mm, and (b) along the well axis. Local increase of SAR in E plane due to TC is shown in grey.

SAR was also computed at different heights, namely z = 0.1 mm and z = 0.37 mm corresponding to the drop of SAR by 10% and 50%, respectively, along the well axis. Results are shown in the Figure 2.22 and compared to the maximum SAR computed at z = 0.02 mm. Peak SAR in the center of the well (distance = 0 mm) decreases by 62% and 12% in respect to the maximum value along E- 114

CHAPTER 2. Electromagnetic and thermal dosimetry at millimeter waves: effects of thermal convection during in vitro exposure of cells and H-planes, respectively, for the corresponding SAR decrease along z by 50% (z = 0.1 mm) and 10% (z = 0.37 mm).

(a) (b) Figure 2.22: Computed SAR without TC at different heights along the well axis for a) E- and b) H- planes.

The impact of the presence of a thin TC on the SAR and temperature distributions within the liquid has been calculated numerically by using the electromagnetic and thermal modules of CST. A 0.3 mm diameter (measured dimension) sphere simulated the TC tip, while a rectangular block of a square cross-section equivalent to the lead in terms of the area (A = 5625 μm2) and length of 10 mm simulated the TC lead, set normal to the E-field. The mesh adopted for the electromagnetic simulation ensured that the TC discretization spanned its smallest dimension with at least 8 mesh cells. The model with the TC presented about 50 millions of mesh cells. Results showed that the introduction of a small TC results in a local increase of SAR close to the TC tip (in E-plane) by about 5% (Fig. 2.23) at z = 0.02 mm. At z = 0.15 mm (i.e., the radius of the TC), SAR distortion around the TC increases (Fig.2.23), and it is about 2 times higher compared to the SAR obtained without TC.

115

CHAPTER 2. Electromagnetic and thermal dosimetry at millimeter waves: effects of thermal convection during in vitro exposure of cells

Figure 2.23: Computed SAR with TC at different heights along the E-plane.

SAR distributions computed in the center of the well (in a 6 mm2 area) with and without TC, at different heights, are compared in Figure 2.24.

(a) (b)

116

CHAPTER 2. Electromagnetic and thermal dosimetry at millimeter waves: effects of thermal convection during in vitro exposure of cells

(c) (d) Figure 2.24: SAR distribution in the center of the well with and without TC in the x, y plane at z = 0.02 mm (a,b) and z = 0.15 mm (c,d). Figures on the left represent the simulations with TC, while figures on the right refer to the same simulations without TC.

Results of electromagnetic simulation have been used as a source for the thermal simulation. Thermal parameters used in the simulation are represented in the Table 2.2. In addition, thermal conductivity k and specific heat capacity cp of polystyrene were 1150 (J/(kg∙°C)) and 0.173 (W/(m∙°C)), respectively [7]. Open boundary conditions are imposed in all the directions. Heat loss toward external environment is controlled in the numerical model by the emissivity and the convection heat transfer coefficient of polystyrene and of the liquid models. The emissivity value was fixed at 0.97 for all the materials. Heat transfer coefficient was set to 19 and 100 for the polystyrene [54] and the liquids [13], respectively. Temperature distributions obtained from thermal simulations are presented in Figure 2.25 in the x,y plane at different heights, including z = 0.02 mm and z = 0.15 mm. They qualitatively show no difference between the configurations with and without the TC.

117

CHAPTER 2. Electromagnetic and thermal dosimetry at millimeter waves: effects of thermal convection during in vitro exposure of cells

(a) (b)

(c) (d) Figure 2.25: Temperature distribution in the center of the well with and without TC in the x, y plane at z=0.02 mm (a, b) and z = 0.15 mm (c, d). Figures on the left represent the simulation with TC, while figures on the right refer to the same simulations without TC.

In particular, temperature distribution computed along x (direction of the peak SAR introduced by the TC) shows that the presence of a small TC set perpendicular to the E-field does not produce artefacts larger than 1.6% when measuring heating in points around the TC. This difference is only negligibly impacted when considering different heights along the well axis (Fig. 2.26). Results of our study confirmed that the heating of the sample will be the same with and without the TC, i.e., TC can be used to correctly measure the temperature elevation induced in SUTs by MMW exposure [46], [47]. Hence, the error in the temperature measurement introduced by a small TC is not due to the distortion of the SAR distribution near the TC but most likely due to the heat transfer within the TC and liquid as detailed in section 2.7 (Temperature-based retrieval of SAR).

118

CHAPTER 2. Electromagnetic and thermal dosimetry at millimeter waves: effects of thermal convection during in vitro exposure of cells

Figure 2.26: Computed temperature distribution with and without the TC along x (direction of the peak SAR introduced by the TC) with and without TC at z = 0.02 mm and z = 0.15 mm.

Finally, we also calculated the reflection coefficient R for water, culture medium, and agar gel using standard equation [55]:

R= (2.13)

where n0 is the refractive index of air equal to 1 and ni is the refractive index of media given by:

(2.14) and

(2.15)

where relative permittivity and conductivity of the media are given in Table 2.1,  = 2f, and f=58.4∙109 Hz. These simple calculations demonstrate that R for water, culture medium, and agar gel was 0.49, 0.48, and 0.49, respectively. No notable differences in reflectivity of these media was observed. Besides, as the thermal properties of these media (Table 2.2) are approximately the same we do not expect that the calibration of TC could be changed by the different SUTs used in the study.

119

CHAPTER 2. Electromagnetic and thermal dosimetry at millimeter waves: effects of thermal convection during in vitro exposure of cells

2.6.1 Impact of cell monolayer The impact of a cell monolayer on the SAR distribution, the amplitude of the electric field, and the power density of the MMW was analytically and numerically investigated in detail at 42 and 58.4 GHz in [56] and [7], respectively. As the exposure conditions are the same as those of this study, results showed below can be applied to the study presented in this work. The exposure scenario is represented by a thin cell monolayer attached to the bottom of the plastic vessel (i.e., a well of a multi-well TCP), covered by a saline solution used to simulate the nutrient medium, and exposed to a plane wave by the bottom of the well as schematically represented in Figure 2.27 [56]. The thickness of the cell monolayer is assumed equal to the average diameter of the cells and it may typically range approximately from 3 to 10 µm. This corresponds to very tightly packed cells, while in experiments typical density of cells is even smaller. The permittivity of the cell monolayer is extrapolated by using the Debye model described in [57] taking into account that at MMW the dielectric permittivity of biological tissues is mainly determined by the free water content (section 1.1.2, Electromagnetic properties). Considering that the epidermis contains 70% water in average [58], both the dielectric permittivity and electric conductivity estimated for the cell monolayer was lower than the cell culture medium (i.e., relative permittivity ε and conductivity σ of cells were 7.19 and 32.94 S/m, respectively [7]). An incident plane wave is partially reflected from the bottom of the well. The transmitted power reaches the cell layer, and then is completely absorbed in the liquid.

Figure 2.27: Model used for the calculation of the SAR and E field exposed to MMW.

In [56] SAR was calculated analytically as:

120

CHAPTER 2. Electromagnetic and thermal dosimetry at millimeter waves: effects of thermal convection during in vitro exposure of cells

SAR = (2.16)

Results obtained for a 42 GHz plane wave showed that the SAR in the cell monolayer is almost two times lower than in the culture medium and it changes only negligibly when increasing the thickness from 5 to 10 µm (Fig. 2.28a). Results are in agreements with the numerical simulation performed in [7] at 58.4 GHz with the finite-difference time-domain (FDTD) method using the same CAD model showed in Figure 2.20. In contrast to the SAR, the E-field does not show a gap in the layer of the cells. The presence of the cells resulted in only a slight increase in the amplitude of the electric field in the medium (solid line of Fig. 2.28b). Besides, the presence of a 10 µm-thick monolayer of cells resulted in 1% reduction of the total power absorbed in the sample, and decreasing the thickness of cell layer this difference becomes even lower. The lower irradiation power absorption in the cell monolayer is mainly due to the lower conductivity of the cells compared to the solution due to the very similar values of the E-field (Eq. (2.16)). The cell monolayer absorbs approximately 3% (i.e., more than 97% of power is absorbed in the culture medium) of the incoming energy, therefore the SAR distribution in the culture medium is only very slightly impacted. This means that the effect of the cell monolayer both on the temperature dynamics as well as on the initiation of the convective currents is expected to be negligible.

(a) (b) Figure 2.28: (a) SAR distribution and (b) electric field distribution in the well with a bottom thickness of 1.5 mm for cell layers of 5 and 10 μm thickness and in the absence of cells (dashed line) exposed to 42 GHz and IPD = 10 mW/cm2 [56].

121

CHAPTER 2. Electromagnetic and thermal dosimetry at millimeter waves: effects of thermal convection during in vitro exposure of cells

2.7 Temperature–based retrieval of SAR

Numerical results were validated experimentally by using a temperature-based approach [7], [33], [48]. The SAR can be estimated by measuring the initial temperature rise rate in the exposed sample:

SAR = (2.17)

where is the specific heat capacity (J/kg/°C), dT is the temperature rise (°C), and dt is the heating time (s). However, the thermal noise obscuring the initial phase of heating kinetics reduces significantly the accuracy of direct measurement of the initial temperature rise rate. As shown in [48] the most reliable way to determine SAR based on temperature measurements is to fit a thermal model to experimental heating kinetics, and then to calculate the initial temperature rise rate from the model. Temperature rise rate in the initial phase of exposure when convection is negligible is proportional to SAR [48] as shown in Equation (2.17). In our exposure conditions, the temperature elevation at the bottom of the liquid induced by MMW exposure depends on the heat transfer due to thermal conduction to the liquid bulk and well bottom contacting with the heated liquid. The thermal model used to fit the experimental temperature dynamics is based on two dimensional heat transfer equations (HTE) in cylindrical coordinates [7]:

for 0 ≤ z < z0 (2.18)

for z ≥ z0 (2.19)

where z0 is distance from external surface of the polystyrene well bottom to the interface between the well bottom and liquid, i.e., the thickness of the well bottom. Origin of coordinates was set in the center of the external surface of the well bottom. In Equations (2.18) and (2.19), T1 = Tp-T0 and

T2=TS –T0, where Tp is polystyrene temperature, TS is the temperature of the SUT (i.e., water, culture medium, and agar gel), and T0 is initial temperature before exposure; µ1 = ρ /k1, µ2 =

ρ /k2, where ρ, cp, and k are mass density, specific heat, and heat conductivity, respectively. The indexes 1 and 2 refer to polystyrene and the SUT, respectively.

122

CHAPTER 2. Electromagnetic and thermal dosimetry at millimeter waves: effects of thermal convection during in vitro exposure of cells

SAR at the bottom of the liquid has a nearly circular distribution as determined from the numerical electromagnetic simulation. It can be approximated by a Gaussian type function as:

(2.20) where c is the constant characterizing beam size and r is radial distance from the beam axis. The value of c, which corresponds to radius ro of the circular beam was experimentally calculated in [7] for the same experimental conditions of this study, according to [48]. It was defined at a level of 2 -1 SARpeak/e, i.e., cro =1 and ro=1/c, and found to be equal to the average value of 9 cm . It corresponds to a radius of the beam ro = 3.3 mm. The heat deposition Q(z,r,t) from Gaussian type beam exposure is defined as [59]:

for z z ≥ 0 (2.21)

Due to small electrical conductivity of polystyrene (i.e., 0.008 S/m), heat deposition in the well bottom is considered negligible compared to that in liquid. In Equations (2.21), in addition to parameters already defined, z coincides with the beam axis (Fig. 2.20), δ is penetration depth of MMW into the liquid, t is time, and u(t) is unit step function. Initial boundary condition for the Equations (2.18) and (2.19) are:

Ti (z, r, 0)= 0 (2.22)

Ti (z, ,0)= 0 t ≥ 0 (2.23)

T2 ( , r, t)= 0 t ≥ 0 (2.24)

t ≥ 0 (2.25)

t ≥ 0 (2.26)

T1 (z0, r, t) = T2 (0, r, t) t ≥ 0 (2.27)

123

CHAPTER 2. Electromagnetic and thermal dosimetry at millimeter waves: effects of thermal convection during in vitro exposure of cells

t ≥ 0 (2.28)

where i = 1, 2 and h is heat transfer coefficient for describing heat loss from the bottom of the well due to air convection. It was calculated according to [60] and found equal to 1.0 W/m2. For short exposures (5–10 s) for estimation of the initial temperature rise rate, the size of the liquid was considered infinite (boundary condition Eq. (2.24)) because the actual size of liquid height has negligible impact on the initial temperature rise. Thermophysical parameters of water, culture medium, and agar gel are reported in Table 2.2 , while the mass density, specific heat capacity and thermal conductivity of polystyrene were 1000 (kg/m3), 1150 (J/(kg∙°C)), and 0.173 (W/(m∙°C)), respectively. Equations (2.18) and (2.19) were solved numerically using an implicit method described in [59]. Temperature dynamics fitted to the above mentioned thermal model for the first 10–15 s of exposure at 195 mW are reported in Figure 2.29 for liquid (i.e., water) and agar gel. The fitting curves (black solid lines) match closely the experimental data. Measured peak SAR values are given in Table 2.3 as mean ± SD determined from five independent recordings. The SAR differences between water, culture medium, and agar gel are within 3% confirming the numerical results (Table 2.3).

(a) (b) Figure 2.29: Temperature dynamics measured in (a) water and (b) agar gel. Corresponding SAR values are reported in Table 2.3. Black solid lines are the fitting curves.

The SAR determined based on temperature measurements is below the computed values (e.g., 19% less for water). This difference results from the 20% underestimation of the temperature rise with 75 µm TC diameter leads as demonstrated in [7] under the same experimental conditions of this work (Fig. 2.30a). In particular, the temperature elevation induced by short exposure to CW 124

CHAPTER 2. Electromagnetic and thermal dosimetry at millimeter waves: effects of thermal convection during in vitro exposure of cells

MMW, recorded by a thin TC with 25 µm diameter resulted in peak SAR value matching 98.5% the computed SAR. This means that for the same experimental conditions, 25 µm diameter TC may be used to retrieve the SAR without any additional correction (1.5% difference is within the statistical error). Larger diameter of TC leads (i.e., 75 µm) introduces a 19.6% error in the temperature measurement, which has to be considered for the correct estimation of the SAR when it is retrieved experimentally. Note that this correction factor is valid only under the same experimental in vitro conditions and it mainly depends on the exposure beam size [7], while it is independent on the media (i.e., water or culture medium) as soon as the beam size is unchanged. Besides, it has been shown [7] that error in temperature measurements is not due to the high temperature gradients along the beam direction near the bottom of a well as they rapidly disappear for exposure times > 1 s, but it is due to the lateral gradients around the TC leads. TC has a high thermal conductivity about 29 times greater than the SUTs. Therefore, temperature in the TC leads reaches its equilibrium over the diameter of TC very quickly. Initial diameter of heated spot in liquid, which coincides with the diameter of the exposure beam for short exposure durations, is 3.3 mm. TC is heated due to the thermal flux from liquid along the thermal gradient in the lateral direction. This results in a temperature gradient in TC and, due to the higher thermal conductivity of TC, to the greater thermal flux in the lateral direction. This reduces equilibrium temperature in TC. The greater the diameter of TC, the greater is the thermal flux along the TC leads from the position of maximal temperature to the periphery. This may explain underestimation of the temperature in measurement made with increasing TC diameter. On the contrary, with decreasing the diameter of TC, its thermal resistance increases and reduces thermal flux along the leads of TC. The hot spot size increases at longer exposure durations, reducing the temperature gradient in lateral direction. This effect could reduce the error in temperature measurement and explain equalization of temperature readings by TC with different diameters during long exposure as showed in Figure 2.30b.

125

CHAPTER 2. Electromagnetic and thermal dosimetry at millimeter waves: effects of thermal convection during in vitro exposure of cells

(a) (b) Figure 2.30: (a) Temperature kinetics measured at bottom of liquid (1.5 ml) using 25 µm (curve 1) TC, 75 µm (curve 2) TC, and FO probe (curve 3). Solid lines (curves 1 and 2) are fitting curves. (b)Temperature changes during long-term exposure. Temperature was measured with two TCs (25 and 75 µm) and FO sensor at bottom of a well filled with 3 ml of liquid and exposed at 195 mW [7].

2.8 Continous wave In this section the heating dynamics of the four SUTs is investigated under the exposure to CW heating. The analysis is performed by considering 1) heating dynamics of the SUTs during long exposure, 2) the dependence on the SAR level, and 3) the role of the liquid volume.

2.8.1 Heating in different media Figure 2.31 shows the temperature dynamics recorded in the four SUTs for CW exposure at 195 mW for 90 min.

126

CHAPTER 2. Electromagnetic and thermal dosimetry at millimeter waves: effects of thermal convection during in vitro exposure of cells

(a)

(b) (c) Figure 2.31: (a) Temperature dynamics of exposure to CW at 195 mW in SUTs. The SUT volume is 2 ml. (b) Zoom within the first 8 s of exposure. (c) Zoom within the first 10 min of exposure. Time point t = 0 corresponds to the starting of the exposure.

The liquid volume used was 2 ml. At the initial stage (during the first 8 s) all curves demonstrate similar dynamics without notable differences (Fig. 2.31b) confirming the negligible changes of SAR among the SUTs presented in the previous sections (i.e., SAR is proportional to the initial temperature rise rate). After about 70 s of exposure, a crest (temperature peak) appears in the water sample (Fig. 2.31c). This moment precedes the initiation of convection in bulk [12], [61], [62]. We assume that this phenomenon is due to the fact that convective currents push up the warm water at the bottom of the well and bring down the cooler water further from the bottom causing the temporary increase and successive decrease of the temperature. At MMW, similar effect was 127

CHAPTER 2. Electromagnetic and thermal dosimetry at millimeter waves: effects of thermal convection during in vitro exposure of cells observed by Khizhnyak and Ziskin [16] during the exposure of 0.5–3.0 mm thick layers of NaCl solution, in the 53–78 GHz range at the SAR levels up to 80 kW/kg. The complex biphasic temperature process in which the initial temperature rise was followed by an asymptotic temperature drop observed in [16] is represented in Figure 2.32.

Figure 2.32: The heating dynamics of the gel (curve 1) and liquid (curve 3) [16].

Besides, our results showed no noticeable difference (< 1%) at 195 mW between culture medium with and without a cell monolayer (Fig. 2.31a), confirming that the presence of a cell layer practically does not change the temperature rise in the culture medium, as well as it does not impact the initiation of convective currents. Convection in water reduces the temperature rise by 3.34 ± 0.12°C (30 ± 3%) with respect to the agar gel at steady state (mean ± SD in five independent recordings). Temperature rise in the culture medium is higher than in water by 0.7°C (8 ± 2%). However, the difference in SAR between culture medium and water is only 2% (Table 2.3). The latter is related to the fact that the contribution of the ionic conductivity into losses decreases with frequency and becomes almost negligible around 60 GHz; this means that at MMW the main contribution is given by the electrical conductivities of water and culture medium which are very close (within the 10%, Table 2.1). The remaining difference can be attributed to different 1) heat capacity (i.e., a higher power should be delivered to water to obtain the same temperature increment as in the culture medium), and 2) viscosity as discussed hereafter. Natural convection may be predicted by using the Rayleigh and Grashof numbers described by Equations (2.7) and (2.8), respectively. Therefore, they have been calculated to provide deeper understanding of the convective problem arising during the in vitro experiments considered in this work.

128

CHAPTER 2. Electromagnetic and thermal dosimetry at millimeter waves: effects of thermal convection during in vitro exposure of cells

The critical Ra number has been calculated for the specific system considered. As convection was initiated in water when its volume exceeded 1.5 ml at ΔT 3 °C (Fig. 2.38c) the critical Ra number was calculated using the parameters given in Table 2.4 and x = 2.9∙10-3 m at ΔT = 3°C. The height of 1.5 ml of water in the center of a well was measured experimentally. It was different from the calculated one (3.9∙10-3 m) due to the formation of meniscus. For considered MMW exposure conditions the critical Ra number of water was found to be equal to 2.3∙104. In 2 ml of water (i.e., x=5.2∙10-3 m) typically used in the experiments, Ra = 1.3∙105 at ΔT = 3°C. This number substantially exceeds the calculated critical Ra number (1708.8) indicating the occurrence of the natural convection in fluids during MMW exposure used in the experiments. The Gr number was calculated using values provided in Table 2.4 and x = 5.2∙10-3 m. It was found that Gr = 3.1∙104 and 1.3∙105 for water, and Gr = 2.8∙104 and 1.15∙105 for culture medium, at ΔT = 3°C and ΔT = 12.5°C, respectively. The values of the Gr number assume that there is a laminar flow in both fluids (Gr < 1109). It is worth to mention that ΔT elevation increases convection of fluid. The ΔT elevation from 3 to 12.5°C increases the Gr number by 4.2 times. The greater the Gr number, the greater is the fluid movement or convection. The Gr numbers for culture medium are about 10% and 13% lower than those for water at ΔT = 3°C and ΔT = 12.5°C, respectively. Viscosity is the only different parameter for the two liquids in the Gr number (Eq. 2.8). Therefore lower convection in cell medium can be attributed to higher viscosity. This, together with 2% difference in SAR, explains higher temperature rise in the culture medium. Indeed, convection transports heat away from the micro TC located at the bottom of the well, reducing the measured temperature.

Table 2.4. Parameters used for calculation of Ra and Gr

2 g (m/s ) β (1/K) x (m) ΔT (°C) (m²/s) (m²/s) (m2/s) (m2/s) 1.5∙ 9.8 0.003 5.2∙ 3 and12.5 1.52∙ 0.63∙ 0.67∙

Moreover, once the Ra number is known, it is also possible to determine the Nu number and retrieve the convection heat transfer coefficient (W/m²·°C) by using Equations (2.11) and (2.12). As reported in [13], in the case of horizontal plates, in the specific situation of upper surface of hot plate, (i.e., SUTs heated from the bottom), the Nu number may be calculated by using the following formula:

129

CHAPTER 2. Electromagnetic and thermal dosimetry at millimeter waves: effects of thermal convection during in vitro exposure of cells

Nu = 0.54 Ra1/4 (2.29) where 0.54 and ¼ are constants empirically determined. The Nu number calculated for water and culture medium, for the typical experimental volume of 2 ml (x = 5.2∙ mm) was 7.165 and 6.985, respectively, confirming the more effective convection in water compared to culture medium. The correspondent coefficients were 861.21 (W/m²·°C) and 831.515 (W/m²·°C) for water and culture medium, respectively. Finally, the Pr number was also defined by using the Equation (2.10) and found to be equal to 4.2 and 4.4 for water and culture medium, respectively. This number is within the typical range reported for water in the 30–40 °C temperature interval [22].

2.8.2 Dependence on SAR Initiation of convection in the liquid samples was investigated as a function of SAR in 1.8 to 24 kW/kg range. Results obtained by comparing the culture medium dynamics with and without cell layer confirmed that the presence of a thin monolayer of cells does not perturb the heating dynamics in the culture medium (Fig. 2.33). The ΔT between the steady state temperature reached in the medium with and without cells varies linearly as a function of the SAR remaining within an average value of 0.2 °C, which is within the standard deviation of the measurements.

Figure 2.33: Difference between steady state temperature of the culture medium with and without cell monolayer as a function of SAR. Experimental data points are represented by grey circles. Each point corresponds to the difference between culture medium with cells and culture medium without cells averaged over three measurements. Linear regression of these points is represented by the solid line.

130

CHAPTER 2. Electromagnetic and thermal dosimetry at millimeter waves: effects of thermal convection during in vitro exposure of cells

Figure 2.34 shows that at SAR ≥ 7.8 kW/kg a temperature peak is generated in the water sample. In the culture medium, the peak appears only at SAR = 24 kW/kg. This can be attributed to higher viscosity of the culture medium compared to water. The occurrence of convection is facilitated by a higher intensity of the source of heating exhibiting a linear relationship to the amplitude of the temperature rise proportional to the SAR [63]. The time point at which the peak appears depends on the SAR. For instance, in water at 7.8 kW/kg convection starts after about one min, while at 24 kW/kg it is triggered after 40 s. Results are in agreement with literature; previous studies demonstrated that the onset of convection occurs as earlier as the heat flux, which can be considered proportional to the SAR, increases [64]–[66].

(a) (b) Figure 2.34: Temperature dynamics for different SAR conditions for (a) water and (b) culture medium. Time point t = 0 corresponds to the beginning of the exposure.

We empirically demonstrated that product of the time point at which convection appears and the temperature reached at the same time point remains constant (55 ± 5 °C∙s). The analysis was performed by exposing water and agar gel to different SAR levels for short duration between 30 and 60 s (Fig. 2.35). The values found were respectively 48.5 ± 6.37 °C∙s, 60 ± 1.2 °C∙s, 53.5 ± 6.48 °C∙s, 57.53 ± 4.45 °C∙s, and 55.18 ± 2.56 °C∙s for SAR corresponding to 3.9 kW/kg, 7.8 kW/kg, 12 kW/kg, 15.9 kW/kg, and 20.5 kW/kg.

131

CHAPTER 2. Electromagnetic and thermal dosimetry at millimeter waves: effects of thermal convection during in vitro exposure of cells

Figure 2.35: Onset of thermal convection as a function of the SAR during short exposure of water and agar gel. Beginning of convection is indicated with a red arrow.

The difference in the temperature dynamics between the cell culture medium and water was also investigated as a function of the SAR in the 1.8– 20.5 kW/kg range (Fig. 2.36).

Figure 2.36: Temperature dynamics in water and culture medium as a function of SAR.

We observed that ΔT increases as a function of SAR, and ΔT(SAR) can be interpolated using a linear regression expressed as ΔT = 0.053∙SAR with R = 0.88 as shown in the Figure 2.37.

132

CHAPTER 2. Electromagnetic and thermal dosimetry at millimeter waves: effects of thermal convection during in vitro exposure of cells

Figure 2.37: Difference between steady state temperature of the culture medium and water as a function of SAR. Experimental data points are represented by white circles. Each point corresponds to the difference between water and culture medium averaged over three measurements. Linear regression of these points is represented by the solid line. Dashed lines indicate the 95% confidence interval.

2.8.3 Role of liquid volume Volume of a liquid sample represents an important parameter to consider when planning an in vitro experiment. In this section, the heating dynamics in water and agar gel was studied by varying the liquid volume from 0.5 to 4 ml corresponding to 1.3 and 10.4 mm height, respectively. Results are illustrated in Figure 2.38. They refer to 30 s exposure to CW at 195 mW.

(a) 0.5 ml (b) 1 ml

133

CHAPTER 2. Electromagnetic and thermal dosimetry at millimeter waves: effects of thermal convection during in vitro exposure of cells

(e) 3 ml (f) 4 ml Figure 2.38: Impact of the liquid volume on triggering convection, maximum heating, and cooling dynamics. The results were obtained for WG input power equal to 195 mW.

For 0.5 ml the temperature rise is significantly higher compared to the volumes exceeding 1 ml. Temperature dynamics of the exposed samples is identical when volume ranges from 1 to 1.5 ml. The effect of convection impacts the temperature dynamics when the volume exceeds 1.5 ml. The greater the volume the earlier the temperature decrease induced by convection is observed (after 24 s, 17 s, 14 s, and 12 s for 1.5 ml, 2 ml, 3 ml, and 4 ml, respectively). The difference between the maximum temperature in the liquid and agar gel also increases with volume (by 5%, 21%, 33%, and 45%, respectively). Note that convection also affects the cooling dynamics, i.e., cooling is faster in larger volumes where the effect of convection is more pronounced. The impact of the liquid volume on the variation of the Gr number has been evaluated for x varying in the range of 1.310-3–10.410-3 m (i.e., the liquid volumes used in experiments). The meniscus effect was not included in our calculation. Besides it is worth to mentions that the

134

CHAPTER 2. Electromagnetic and thermal dosimetry at millimeter waves: effects of thermal convection during in vitro exposure of cells

meniscus effect considered as an important factor for frequencies around 1 GHz [67], [68] can be neglected in bioelectromagnetic studies at MMW due to the shallow penetration. The results of calculations obtained using Equation (2.8) and parameters for g, β, ΔT, and ν given in Table 2.4 are presented in Table 2.5. The increase of ΔT and x results in elevation of Gr number. It is interesting to observe that the effect of changing the liquid volume is much higher than the one occurred by changing the temperature. The results show that an increase of the liquid volume from 0.5 to 4 ml results in a rise of the Gr number of about three orders of magnitude explaining why with increasing the volume of fluid the maximum temperature elevation was reduced and cooling was faster, i.e., the effect was due to increased convection. These results suggest that the culture medium volume in in vitro experiments at MMW has to be carefully chosen due to convection, which impacts the maximum temperature as well as the cooling dynamics.

Table 2.5. Dependence of the Gr number on the temperature difference ΔT and the height of liquid x

ΔT (°C) 3 12.5 V (ml) 0.5 1 1.5 2 4 0.5 1 1.5 2 4 1.3 2.6 3.9 5.2 10.4 1.3 2.6 3.9 5.2 10.4 x (10-3 m)

4.9∙102 3.9∙103 1.3∙104 3.1∙104 2.5∙105 2.0∙103 1.6∙104 5.5∙104 1.3∙105 1.0∙106 Grwater

Gr culture 4.3∙102 3.45∙103 1.2∙104 2.8∙104 2.2∙105 1.8∙103 1.4∙104 4.9∙104 1.15∙105 9.1∙105 medium

Similar effect is observed in Figure 2.15 taken from [7] where the lower baseline temperature recorded in greater volumes of liquid resulted from the enhanced cooling effect of convection.

2.9 Pulse-induced heating The appearance of the convective currents has been evaluated 1) as a function of the pulse duration for single pulse, and 2) for long exposure of the SUTs to 1.5 s pulses.

2.9.1 Dependence on pulse duration

The initiation of convection is investigated for pulses with a duration τ ranging from 1 to 6.5 s (Fig. 2.39). Hereafter all results are presented for the water and agar gel volumes equal to 2 ml. 135

CHAPTER 2. Electromagnetic and thermal dosimetry at millimeter waves: effects of thermal convection during in vitro exposure of cells

Temperature pulses present an asymmetrical shape, characterized by the fast increase of temperature after application of pulse-modulated MMW followed by its slow drop after ending the pulse. The passive cooling of the heating SUT determines the progressive temperature drop. For the constant peak power of 4.2 W, the amplitude of the temperature pulses increases with the pulse duration from 8 (1 s pulse) to 21.4 °C (6.5 s pulse) and from 8 (1 s pulse) to 25 °C (6.5 s pulse)in water and gel, respectively (temperature rise is proportional to SAR). When convection is absent (τ = 1 s) the shape of the temperature pulses for both samples is identical. For τ = 1.5, the amplitude of the temperature pulse remains the same. However, the cooling rate in the agar gel is slightly faster in the time interval from 0 to 3 s after exposure, and afterwards it becomes slower compared to that in the liquid. For τ = 2 to 4 s, the amplitude of the temperature pulse generated in water is higher than the one in agar gel. The maximum difference (2.13 ± 0.57 °C) occurs for τ = 3 s. We presume that this phenomenon is of the same nature as for the temperature peak presented in Figure 5c. The time difference of appearance of this peak at CW (70 s) and PW (3 s) exposures is caused by the difference in power (195 mW for CW and 4.2 W for PW). For τ ≥ 4.5 s, convection strongly changes the heat pulse shape decreasing the peak temperature. In addition, it increases the cooling rate from 1.8 to 7.12 °C/s when pulse duration increases from 1 to 6.5 s (here the cooling rate is averaged over the initial cooling phase till the temperature drops to ΔTmax/2). These results suggest that changes in the pulse shape due to convection depend on the pulse duration. Convective currents during the warm up phase can be avoided or reduced by decreasing the pulse duration.

(a) 1 s (b) 1.5 s

136

CHAPTER 2. Electromagnetic and thermal dosimetry at millimeter waves: effects of thermal convection during in vitro exposure of cells

(e) 3 s (f) 3.5 s

(g) 4 s (h) 4.5 s

137

CHAPTER 2. Electromagnetic and thermal dosimetry at millimeter waves: effects of thermal convection during in vitro exposure of cells

(i) 5 s (j) 5.5 s

(k) 6 s (l) 6.5 s

Figure 2.39: Heating dynamics in water and agar gel for different pulse durations between 1–6.5 s.

2.9.2 Heating in water and agar gel

Water and agar gel were exposed for 90 min to the electromagnetic pulse train consisted of 1.5 s square-wave amplitude-modulated MMW pulses with a peak power of 4.2 W and period of 20 s. This corresponds to the pulse duration which is impacted by the thermal convection only slightly in the cooling phase (Fig. 2.39b). Measured heat pulses in the SUTs have an initial amplitude of 10 °C (Fig. 2.40a). A decrease of the peak amplitude by 9% and 11% in water and agar gel, respectively, probably caused by the slow heat transfer to the remote wall of a well, was observed in both samples after 10 min of exposure from the beginning until the end of exposure (Fig. 2.40b). While the amplitude of the peak remains constant after 10 min of exposure for both samples, the cooling dynamics of heat pulses with and without convection are different. They remain similar for about 3 s (the cooling is slightly faster in the ager gel). Afterwards, convection in water results in a faster 138

CHAPTER 2. Electromagnetic and thermal dosimetry at millimeter waves: effects of thermal convection during in vitro exposure of cells decrease of temperature, leading to the maximum difference of 1.3°C at the end of the cooling phase of a pulse compared to the agar gel. This is responsible for the lower average temperature in the liquid compared to the agar gel (Figs. 2.40c,d). Average temperature (black line in Figure 2.40) has been calculated by using a moving average filter with 75 s span by means of the smooth function of Matlab [69]. The filter smooths data by replacing each data point with the average of the neighboring data points defined within the span. These results suggest that, for short pulses, the main effect of convection is a decrease of the average temperature resulting from increased cooling rate in liquids compared to the agar gel.

(a) (b)

(c) (d) Figure 2.40: Temperature dynamics in PW regime in water (c) and agar gel (d). Black lines represent the average temperature calculated with 75 s span. Subplots on the top compare the dynamics in water with the one in the agar at the beginning of the exposure (a) and at the end (b).

139

CHAPTER 2. Electromagnetic and thermal dosimetry at millimeter waves: effects of thermal convection during in vitro exposure of cells

2.10 Conclusions

In this chapter, electromagnetic and thermal dosimetry at MMW in a specific in vitro exposure scenario represented by one well of a 12-well TCP exposed from the bottom by an open-ended rectangular WG at 58.4 GHz is presented. The methodology used to measure the temperature by means of thin TC is detailed analysing the possible issues deriving from the incorrect alignment of the TC in respect to the E-field. Afterwards, numerical results of peak SAR values are compared to experimental results of SAR determined from the temperature measurements. Finally, the effects of convection on temperature dynamics in models representing a typical in vitro exposure scenario at MMW in CW and PW regimes are detailed. Temperature dynamics in four different samples were investigated, i.e., distilled water, a cell monolayer attached to the bottom of a 12-well TCP and covered by the culture medium, culture medium without cells, and convection-free medium with electromagnetic properties close to water (e.g., agar gel). It has been experimentally demonstrated that a cell monolayer practically does not change heating dynamics confirming the analytical results of previous studies (i.e., the presence of a thin cell monolayer only slightly perturbs the power absorption in in vitro samples at MMW). This suggests that when TC is used to monitor the local temperature dynamics with a microscale resolution, the measurements can be performed in a separate well without cells to avoid damaging and contamination of cells. For CW exposure, a temperature peak appears after several tens of seconds, preceding initiation of convection in bulk. The time point at which the peak appears depends on SAR. The higher the SAR the earlier the peak appears triggering convection. It is more pronounced in free water compared to the culture medium due to lower viscosity of water. This effect has to be carefully taken into account in in vitro bioelectromagnetic experiments as it may impact the biological outcome and result in a different cellular response compared to conventional heating used as a control. Our results also indicate that the liquid volume is an important parameter impacting convection, which effect is often ignored or underestimated in in vitro experiments. Indeed, convection is triggered earlier when the volume increases. Increasing the volume also results in a greater cooling effect, impacting temperature dynamics at the cellular level. Furthermore, it also affects the cooling dynamics after exposure switched off, i.e., the cooling is faster in larger volumes. In PW regime, convection strongly depends on the pulse duration. For the exposures in these experiments, at τ = 1 s, the shape of the thermal pulse is not impacted by convection. For τ = 1.5, the amplitude of the thermal pulses remains the same. However, the cooling dynamics is slightly 140

CHAPTER 2. Electromagnetic and thermal dosimetry at millimeter waves: effects of thermal convection during in vitro exposure of cells affected by convection resulting in a decrease of the average temperature for prolonged exposures. For τ = 2 to 4 s, convection increases the amplitude of the thermal pulse in liquid, and this effect is the most pronounced for τ = 3 s for the considered exposure condition. The exact understanding of the mechanism of the observed effect would require analysis of convective currents in 3D, which is not achievable using the techniques employed in this study. Detailed analysis of this effect is a future aim of our group. For τ ≥ 4.5 s, convection strongly changes the heat pulse shape decreasing the peak temperature and increasing the cooling rate. Note that heating was studied under the conditions of local heating with narrow beam using open-ended waveguide. When samples are exposed with antennas creating a broader beam the greater areas of uniform heating at the bottom of the well can occur. In this case the boundaries for convection will also change, potentially affecting the resulting heating. Results presented in this work suggest that in planning in vitro studies at MMW convection is one of the parameters that has to be carefully taken into account. It may impact cellular response through indirect mechanisms, such as local change of the concentration of oxygen or nutrients transported by the culture medium. Deeper investigation of convection mechanisms requires numerical analysis of multi-physics problem, which constitutes one of the perspectives of the current study as detailed in the section Future work and perspectives.

141

CHAPTER 2. Bibliography

Bibliography

[1] A. G. Pakhomov, Y. Akyel, O. N. Pakhomova, B. E. Stuck, and M. R. Murphy, “Current state and implications of research on biological effects of millimeter waves: a review of the literature,” Bioelectromagnetics, vol. 19, no. 7, pp. 393–413, 1998. [2] Y. Le Dréan et al., “State of knowledge on biological effects at 40–60 GHz,” Comptes Rendus Phys., vol. 14, no. 5, pp. 402–411, May 2013. [3] M. Cueille, A. Collin, C. Pivain, and P. Leveque, “Development of a numerical model connecting electromagnetism, thermal and hydrodynamics to analyse in vitro exposure system,” Ann. Telecommun. - Ann. Télécommunications, vol. 63, no. 1, pp. 17–28, Feb. 2008. [4] A. Paffi, M. Liberti, F. Apollonio, A. Sheppard, and Q. Balzano, “In vitro exposure: linear and non-linear thermodynamic events in petri dishes,” Bioelectromagnetics, vol. 36, no. 7, pp. 527– 537, Oct. 2015. [5] S. M. Lindsay and J. Yin, “Temperature gradients drive radial fluid flow in petri dishes and multiwell plates,” AIChE J., vol. 62, no. 6, pp. 2227–2233, Jun. 2016. [6] J. X. Zhao, “Numerical dosimetry for cells under millimetre-wave irradiation using petri dish exposure set-ups,” Phys. Med. Biol., vol. 50, no. 14, p. 3405, 2005. [7] M. Zhadobov, S. I. Alekseev, R. Sauleau, Y. Le Page, Y. Le Dréan, and E. E. Fesenko, “Microscale temperature and SAR measurements in cell monolayer models exposed to millimeter waves,” Bioelectromagnetics, vol. 38, no. 1, pp. 11–21, Jan. 2017. [8] A. V. Boriskin et al., “Enhancing exposure efficiency and uniformity using a choke ring antenna: application to bioelectromagnetic studies at 60 GHz,” IEEE Trans. Microw. Theory Tech., vol. 61, no. 5, pp. 2005–2014, May 2013. [9] A. Guy, C. Chou, and J. McDougall, “A quarter century of in vitro research: a new look at exposure methods,” Bioelectromagnetics, vol. 4, no. 21, pp. 21–39. [10] Q. Zhang, T. H. Jackson, and A. Ungan, “Numerical modeling of microwave induced natural convection,” Int. J. Heat Mass Transf., vol. 43, no. 12, pp. 2141–2154, Jun. 2000. [11] K. G. Ayappa, S. Brandon, J. J. Derby, H. T. Davis, and E. A. Davis, “Microwave driven convection in a square cavity,” Am. Inst. Chem. Eng. J., vol. 40, no. 7, pp. 1268–1272, 1994. [12] I. Tanasawa, “Experimental techniques in natural convection,” Exp. Therm. Fluid Sci., vol. 10, no. 4, pp. 503–518, May 1995. [13] T. L. Bergman, A. S. Lavine, F. P. Incropera, and D. P. Dewitt, Fundamentals of heat and mass transfer, Seventh Edn. Hoboken, NJ: John Wiley, 2007. [14] O. V. Betskii, N. D. Devyatkov, and V. V. Kislov, “Low intensity millimeter waves in medicine and biology,” Crit. Rev. Biomed. Eng., vol. 28, no. 1–2, pp. 247–268, 2000. [15] V. S. Sharov, K. D. Kazarinov, V. E. Andreev, A. V. Putvinskiĭ, and O. V. Betskiĭ, “[Acceleration of peroxide oxidation of lipids under electromagnetic radiation in the millimeter range],” Biofizika, vol. 28, no. 1, pp. 146–148, Feb. 1983. [16] E. P. Khizhnyak and M. C. Ziskin, “Temperature oscillations in liquid media caused by continuous (nonmodulated) millimeter wavelength electromagnetic irradiation,” Bioelectromagnetics, vol. 17, no. 3, pp. 223–229, 1996. [17] W.F. Pickard, W. L. Straube, E. G. Moros, and X. Fan, “Simplified model and measurement of specific absorption rate distribution in a culture flask within a transverse electromagnetic mode exposure system,” Bioelectromagnetics, vol. 20, no. 3, pp. 183–93, 1999. [18] J. Sandblom and S. Theander, “The effect of microwave radiation on the stability and formation of gramicidin‐A channels in lipid bilayer membranes,” Bioelectromagnetics, vol. 12, no. 1, pp. 9–20, 1991. [19] A. Paffi, F. Feo, M. Liberti, F. Apollonio, A. Sheppard, and Q. Balzano, “Vortical motion of liquid medium in small, shallow, containers with different end wall temperatures,” presented at the BioEM 2018, Piran-Portorož, Slovenia, 25-29 June.

142

CHAPTER 2. Bibliography

[20] T. Wessapan and P. Rattanadecho, “Aqueous humor natural convection of the human eye induced by electromagnetic fields: in the supine position,” J. Med. Bioeng., pp. 241–258, 2014. [21] E. P. Khizhnyak and M. C. Ziskin, “Heating patterns in biological tissue phantoms caused by millimeter wave electromagnetic irradiation,” IEEE Trans. Biomed. Eng., vol. 41, no. 9, pp. 865–873, Sep. 1994. [22] R. B. Bird, W. E. Stewart, and E. N. Lightfoot, Transport Phenomena. John Wiley & Sons, 2007. [23] A. J. Chapman, Heat Transfer. Macmillan Publishers Limited, 1984. [24] J. H. I. Lienhard and Lienhard J.H .V, in A heat transfer textbook, fourth edition, Fourth Edn., Cambridge, MA: Phlogiston Press, 1981, pp. 401–455. [25] K. M. Sekins and A. F. Emery, “Thermal science for physical medicine,” in Therapeutic heat and cold, Third Edn., Baltimore, MD: Lehmann JF : Williams and Wilkins, 1982, pp. 70– 132. [26] W. M. Rohsenow, J. P. Hartnett, and Y. I. Cho, Handbook of heat transfer. McGraw-Hill, 1998. [27] Y. A. Çengel, Heat Transfer: A Practical Approach. McGraw-Hill, 2003. [28] A. Bejan, Convection Heat Transfer, Fourth Edn. New York, NY: John Wiley & Sons. [29] H. Jeffreys, “LXXVI. The stability of a layer of fluid heated below,” Lond. Edinb. Dublin Philos. Mag. J. Sci., vol. 2, no. 10, pp. 833–844. [30] W. L. Heitz and J. W. Westwater, “Critical Rayleigh numbers for natural convection of water confined in square cells with L/D from 0.5 to 8,” J. Heat Transf., vol. 93, no. 2, pp. 188– 195, May 1971. [31] D. J. Giard et al., “In vitro cultivation of human tumors: establishment of cell lines derived from a series of solid tumors,” J. Natl. Cancer Inst., vol. 51, no. 5, pp. 1417–1423, Nov. 1973. [32] C. Le Quément, C. N. Nicolaz, D. Habauzit, M. Zhadobov, R. Sauleau, and Y. Le Dréan, “Impact of 60-GHz millimeter waves and corresponding heat effect on endoplasmic reticulum stress sensor gene expression,” Bioelectromagnetics, vol. 35, no. 6, pp. 444–451, Sep. 2014. [33] N. Chahat, M. Zhadobov, R. Sauleau, and S. I. Alekseev, “New method for determining dielectric properties of skin and phantoms at millimeter waves based on heating kinetics,” IEEE Trans. Microw. Theory Tech., vol. 60, no. 3, pp. 827–832, Mar. 2012. [34] F. Erdogdu, Optimization in food engineering. CRC Press, 2008. [35] W. J. Ellison, “Permittivity of pure water, at standard atmospheric pressure, over the frequency range 0–25 THz and the temperature range 0–100°C,” J. Phys. Chem. Ref. Data, vol. 36, no. 1, pp. 1–18, Feb. 2007. [36] A. R. Guraliuc, M. Zhadobov, O. De Sagazan, and R. Sauleau, “Solid phantom for body- centric propagation measurements at 60 GHz,” IEEE Trans. Microw. Theory Tech., vol. 62, no. 6, pp. 1373–1380, Jun. 2014. [37] S. M. R. Goodhew and R. Griffiths, “Analysis of thermal-probe measurements using an iterative method to give sample conductivity and diffusivity data,” J. Appl. Energy, vol. 77, pp. 205–223, 2004. [38] D. W. Sun, Computational fluid dynamics in food processing. Boca Raton, FL: CRC Press, 2007. [39] M. Zhang et al., “Experimental determination of thermal conductivity of water−agar gel at different concentrations and temperatures,” J. Chem. Eng. Data, vol. 56, no. 4, pp. 859–864, Apr. 2011. [40] IAPWS, “International Association for the Properties of Water and Steam,” Mar-2019. [Online]. Available: http://www.iapws.org/. [41] E. Fröhlich et al., “Comparison of two in vitro systems to assess cellular effects of nanoparticles-containing aerosols,” Toxicol. Vitro Int. J. Publ. Assoc. BIBRA, vol. 27, no. 1, pp. 409–417, Feb. 2013.

143

CHAPTER 2. Bibliography

[42] E. Fernández, D. López, C. Mijangos, M. Duskova‐Smrckova, M. Ilavsky, and K. Dusek, “Rheological and thermal properties of agarose aqueous solutions and hydrogels,” J. Polym. Sci. Part B Polym. Phys., vol. 46, no. 3, pp. 322–328, Feb. 2008. [43] M. Zhadobov, R. Sauleau, R. Augustine, C. Le Quément, Y. Le Dréan, and D. Thouroude, “Near-field dosimetry for in vitro exposure of human cells at 60 GHz,” Bioelectromagnetics, vol. 33, no. 1, pp. 55–64, Jan. 2012. [44] M. Kojima et al., “Characteristics of ocular temperature elevations after exposure to quasi- and millimeter waves (18-40 GHz),” J. Infrared Millim. Terahertz Waves, vol. 36, no. 4, pp. 390–399, Apr. 2015. [45] Keysight Technologies, “Practical Temperature Measurements,” http://literature.cdn.keysight.com/litweb/pdf/5965-7822E.pdf. . [46] S. I. Alekseev and M. C. Ziskin, “Distortion of millimeter-wave absorption in biological media due to presence of thermocouples and other objects,” IEEE Trans. Biomed. Eng., vol. 48, no. 9, pp. 1013–1019, Sep. 2001. [47] S. I. Alekseev, M. C. Ziskin, and E. E. Fesenko, “Problems of using a thermocouple for measurements of skin temperature rise during the exposure to millimeter waves,” Biophysics, vol. 56, no. 3, p. 525, Aug. 2011. [48] S. I. Alekseev and M. C. Ziskin, “Local heating of human skin by millimeter waves: a kinetics study,” Bioelectromagnetics, vol. 24, no. 8, pp. 571–581, Dec. 2003. [49] C. A. Balanis, Antenna Theory: Analysis and Design. John Wiley & Sons, 2016. [50] T. C. Cetas, “Thermometry,” in Therapeutic heat and cold, Third Edn., Baltimore, MD: Lehmann JF : Williams and Wilkins, 1982, pp. 35–69. [51] P. B. Dunscombe, R. T. Constable, and J. McLellan, “Minimizing the self-heating artefacts due to the microwave irradiation of thermocouples,” Int. J. Hyperthermia, vol. 4, no. 4, pp. 437–45, Aug. 1988. [52] P. B. Dunscombe, J. McLellan, and K. Malaker, “Heat production in microwave-irradiated thermocouples,” Med. Phys., vol. 13, no. 4, pp. 457–61, Aug. 1986. [53] R. T. Constable, P. B. Dunscombe, A. Tsoukatos, and K. Malaker, “Perturbation of the temperature distribution in microwave irradiated tissue due to the presence of metallic thermometers,” Med. Phys., vol. 14, no. 3, pp. 385–8, Jun. 1987. [54] R. Conti, A. A. Gallitto, and E. Fiordilino, “Measurement of the convective heat-transfer coefficient,” Phys. Teach., vol. 52, no. 2, pp. 109–111, Feb. 2014. [55] C. Polk and E. Postow, Handbook of Biological Effects of Electromagnetic Fields, Third Edition - 2 Volume Set. CRC Press, 1995. [56] S. I. Alekseev, M. V. Zhadobov, E. E. Fesenko, and E. E. Fesenko, “Millimeter wave dosimetry at exposure of cell monolayers,” Biophysics, vol. 62, no. 2, pp. 261–264, Mar. 2017. [57] S. Alekseev and M. Ziskin, “Human skin permittivity determined by millimeter wave reflection measurements,” Bioelectromagnetics, vol. 28, no. 5, pp. 331–339, Jul. 2007. [58] K. Kageyama, Y. Onoyama, H. Kogawa, E. Goto, and K. Tanabe, “The maximum and minimum water content and cell volume of human erythrocytes in vitro,” Biophys. Chem., vol. 34, no. 1, pp. 79–82, Sep. 1989. [59] S. I. Alekseev, A. A. Radzievsky, I. Szabo, and M. c. Ziskin, “Local heating of human skin by millimeter waves: Effect of blood flow,” Bioelectromagnetics, vol. 26, no. 6, pp. 489–501, Sep. 2005. [60] W. H. McAdams, Heat Transmission. 1954. [61] M. J. Khol, M. Kristoffersen, and F. A. Kulacki, “Stability and convection in impulsively heated porous layers,” J. Heat Transf., vol. 130, no. 11, pp. 112601–9. [62] M. Kaviany, “Onset of thermal convection in a fluid layer subjected to transient heating from below,” J. Heat Transf., vol. 106, no. 4, pp. 817–823, Nov. 1984. [63] B. V. Antohe and J. L. Lage, “Amplitude effect on convection induced by time-periodic horizontal heating,” Int. J. Heat Mass Transf., vol. 39, no. 6, pp. 1121–1133, Apr. 1996. 144

CHAPTER 2. Bibliography

[64] D. J. Yang and C. K. Choi, “The onset of thermal convection in a horizontal fluid layer heated from below with time-dependent heat flux,” Phys. Fluids, vol. 14, pp. 930–937, Mar. 2002. [65] B. S. Jhaveri and G. M. Homsy, “The onset of convection in fluid layers heated rapidly in a time-dependent manner,” J. Fluid Mech., vol. 114, pp. 251–260, Jan. 1982. [66] K.-H. Kim and M.-U. Kim, “The onset of natural convection in a fluid layer suddenly heated from below,” Int. J. Heat Mass Transf., vol. 29, no. 2, pp. 193–201, Feb. 1986. [67] J. Schuderer and N. Kuster, “Effect of the meniscus at the solid/liquid interface on the SAR distribution in Petri dishes and flasks,” Bioelectromagnetics, vol. 24, no. 2, pp. 103–108, Feb. 2003. [68] A. Paffi, F. Apollonio, M. Liberti, and Q. Balzano, “Effect of the meniscus at the solid- liquid interface on the microwave exposure of biological samples,” in 2014 44th European Microwave Conference, 2014, pp. 691–694. [69] MathWorks, “Smooth response data - MATLAB smooth.” [Online]. Available: https://www.mathworks.com/help/curvefit/smooth.html.

145

Chapter 3 Cell mortality and heat shock stress induced by pulsed modulated millimeter wave heating at cellular level

Content

3.1 Introduction...... 147 3.2 Biological models ...... 149

3.3 Electromagnetic and thermal dosimetry ...... 158

3.3.1 1.5 s heat pulses ...... 159 3.3.2 6 s heat pulses ...... 162 3.3.3 Thermal dose ...... 167

3.4 Heat shock response ...... 168

3.4.1 Dose-response curves for continuous wave heating ...... 170 3.4.2 Pulsed vs continuous heating ...... 173

3.5 Cellular apoptosis ...... 178

3.5.1 Theoretical basis of cellular apoptosis ...... 179 3.5.2 Pulsed vs continuous heating ...... 182

3.6 Discussion ...... 189 3.7 Conclusions ...... 192 Bibliography ...... 193

CHAPTER 3. Cell mortality and heat shock stress induced by pulsed modulated millimeter wave heating at cellular level

3.1 Introduction

Different power sources have been used to induce heating of biological tissues, including ultrasound and various bands of the electromagnetic spectrum i.e., radiofrequency (RF), microwave (MW), ultraviolet (photothermal therapy), and light (lasers) as detailed in the Chapter 1 of this thesis (sections in 1.3.1, Hyperthermia - Thermal ablation) [1]. In the MW range, devices used for thermal ablation generally operate in the Industrial Scientific Medical (ISM) bands around 434 MHz, 915 MHz, and 2.45 GHz. Numerous advantages of higher frequencies, such as higher spatial resolution and compact size of radiating structures have motivated exploration of MW up to 18 GHz for local tumor ablation with minimized invasiveness and collateral damages (9.2 GHz [2], 10 GHz [3], 14.5 GHz [4], and 18 GHz [5]). However, the use of the MMW band remains still unexplored for this purpose. It has been recently theoretically demonstrated [6] that the 20–100 GHz range may be employed for high spatially-accurate focusing of heat inside the skin by varying i) frequency (i.e., the higher the exposure frequency, the smaller the penetration depth of heat, and the closer the temperature peak to skin surface), ii) exposure beam size (i.e., larger exposure beam size results in a decrease of skin surface temperature and elevation of temperature of deeper tissue layers), and iii) air convection (i.e., enforcement of air convection reduces overheating of skin surface shifting the maximum heating towards deeper skin layers. Note that higher spatial resolution can be achieved in this band and locally higher absolute temperatures and gradients can be induced due to increasing with frequency transmission at air/tissue interface and more localized absorption. Compared to laser radiations, MMW range allows to shift the maximal heating towards deeper skin layers. These observations suggest that the lower part of the MMW range constitutes a promising alternative for non-invasive, spatially- accurate, and controllable thermal treatment of superficial skin cancers such as spreading melanoma (section 1.1.1, Skin cancer), representing a very aggressive tumour extremely resistant to both chemotherapy and radiotherapy [7] located within the MMW penetration depth. This potential technique has certain advantages over other hyperthermia sources: 1) it is a non-invasive treatment, 2) the maximal temperature desired may be precisely regulated within the skin at the melanoma region by adjusting exposure frequency, beam size and shape, and airflow velocity, 3) temperature elevation of skin surface may be monitored using an infrared camera or other temperature measurement devices, 4) forced air convection enables to avoid burning of the epidermis and overheating of deeper tissues at higher exposure intensities, and 5) radiating structures are more compact compared to those at lower MW frequencies allowing to selectively target the melanoma region without overheating neighbouring healthy tissues. 147

CHAPTER 3. Cell mortality and heat shock stress induced by pulsed modulated millimeter wave heating at cellular level

The biological rationale for the application of heat to treat melanoma is mediated by the high susceptibility of this type of cancer to elevated temperatures. Several studies performed both in vivo and in vitro have demonstrated the efficacy of heat to treat melanoma alone [8] or in combination of other therapies, including magnetic nanoparticles such as cationic lyposomes [9], [10], chemotherapy [11], and radiotherapy [12], [13]. In particular, Szabo et al., [14] demonstrated that temperature elevation within the 39–53 °C temperature interval, induced by 42.25 GHz MMW at incident power density between 0.74 to 1.48 W/cm2, revealed lower survival rate of melanoma cells than keratinocytes (predominant healthy cells of the skin). Induction of cell death as a response to hyperthermia in malignant melanoma cells was investigated in several studies. In vitro experiments performed on melanoma cells have demonstrated a variety of effects for heating temperatures ranging from 41–48 °C including the reduction of cell viability in a time and temperature-dependent manner [15], the activation of an apoptotic pathway [13] or the endoplasmic reticulum (ER) stress, and ER-mediated apoptosis [16]. It has been recently demonstrated that hyperthermic temperatures (i.e., 43°C to 45°C) are able to trigger both extrinsic and intrinsic apoptotic pathways in melanoma cells [13] mediated by the activation of Caspases, a family of protease enzymes playing essential roles in programmed cell death [17]. They are synthesized as inactive precursor molecules (pro- Caspase) and are converted by proteolytic cleavage to active enzymes. In particular, it was shown [13] that incubation of melanoma cells at 45°C for 2 h induced activation of Caspase- (Casp-) 3, -6, and -7 up to 24 h (-3) and 72 h (-6,-7) post exposure. All the above-mentioned studies exploited the cytotoxic effect of heat delivered in a continuous manner. However, in contrast to conventional hyperthermia, pulsed electromagnetically-induced heating can lead to stronger damage in cells compared to continuous heating [18]. Short thermal pulses provide the rapid achievement of high peak temperature in the area of interest, eluding the injury of the healthy tissue surrounding the tumor. Moreover, heat pulses may provide other advantages [19], [20] (section 1.3.3, Importance of pulsed waveform in heat-induced cellular responses): i) to decrease treatment duration reducing patient discomfort and overall costs of the treatment, ii) to eliminate or reduce the influence of blood circulation, and iii) to eliminate or reduce the development of cellular thermotolerance. As detailed in the Chapter 1 (sub-section in 1.3.1, Cellular and molecular response to heat), depending on parameters and conditions of heating, cells may respond to temperature elevation in different ways. In case of heat shock at sub-lethal temperatures (typically ranging from 39 to 42°C), thermotolerance is the dominant cellular response [21]. In this case, cells overexpress chaperones and heat shock proteins (HSP) that protect cellular proteins from misfolding and aggregation [22]. Various forms of cellular stress (e.g., moderate heat) dramatically increase the phosphorylation of 148

CHAPTER 3. Cell mortality and heat shock stress induced by pulsed modulated millimeter wave heating at cellular level small molecular weight chaperone such as HSP27 [8], which promotes recognition of client proteins through an oligomerization phenomenon. On the contrary, when cells undergo to sever heat shock, typically > 43°C, the HSP response may be not able to cope with thermal stress, leading to cellular apoptosis. Most of the studies dealing with pulsed heating have been performed so far to analyze the thermal damage mainly in relation to ablation at temperatures exceeding the lethal threshold [23], [24]. Therefore, the exact role of heat pulse parameters at sub-lethal average temperatures in terms of heat shock response remains unclear. Nevertheless, as the interest to pulsed thermal treatment of tumors is increasing, the study of the cellular response, both in terms of activation of cellular repair processes mediated by HSP induction and heat-induced damage is of importance. The main objective of this study is to compare the response of malignant melanoma cells following exposure to continuous (CW) and pulse-modulated (PW) MMW-induced heating which, for the best of our knowledge, has never been investigated so far. First, heat pulses were locally generated at the cellular level in vitro using an ad hoc MMW exposure system. Second, the heat shock response was quantified by following the phosphorylation of HSP27 as a marker of heat-induced cellular stress for the continuous and pulsed heating with the same average temperature rise. Finally, Casp-3 cleaved activation was evaluated as a marker of cellular damage in cells. Different durations of pulses were analysed while maintaining the same peak and average temperatures. We chose an experimental approach based on fluorescence microscopy image analysis as it permits to follow in space the extent of the biological response, to quantify the spatially accurate focusing, and to evaluate effects of exposure at the periphery of the beam. The chapter is organized as follows. First, the biological models including the cell line, exposure protocol, picture acquisition, data analysis, and thermal dose quantification are presented. Second, electromagnetic and thermal dosimetry techniques are described. Third, the heat shock response mediated by the phosphorylation of HSP27 is presented; theoretical basis are firstly provided followed by the presentation and analysis of the results obtained. Fourth, induced cellular damage mediated by the activation of cleaved Casp-3 is presented; theoretical basis of the extrinsic and intrinsic mechanisms of cellular apoptosis and analysis of the results are provided. Finally, the main discussion and conclusions are drawn.

3.2 Biological models

149

CHAPTER 3. Cell mortality and heat shock stress induced by pulsed modulated millimeter wave heating at cellular level

Previous studies demonstrated that cellular response to heat largely varies, depending on the cell line, temperature, and exposure time [15], [25]. Numerous studies aimed to investigate the cellular response of several melanoma cell lines by using different experimental protocols [13], [15], [26]– [28]. However, to the best of our knowledge, the cellular response of the A375 melanoma cell line to pulsed heating has never been investigated in detail, representing the aim of this thesis. In this study, melanoma cells were cultured in a standard 12-well tissue culture plate made of polystyrene and exposed from the bottom by an open-ended rectangular WR15 waveguide (WG) antenna located 5 mm from the plate inside the incubator.

3.2.1 Cell line The human malignant melanoma A375 cells were purchased from American Type Culture Collection (ATCC, Molsheim, FR). These cells were cultured in Dulbecco’s modified Eagle medium (Gibco/Life Technologies, Carlsbad, CA, USA) supplemented with 8% fetal calf serum

(FCS), 1% antibiotics, and 1% L-glutamine, in a humidified incubator at 37 °C and 5% CO2. Cells were grown as monolayer culture and the medium was renewed every 2 days. When cells reached 70 to 80% confluence, they are passaged in order to maintain their proliferation phenotype. To avoid any problem of senescence or drift of the cellular population, the experiments were conducted at earlier passages (between 4 and 10).

(a) (b)

150

CHAPTER 3. Cell mortality and heat shock stress induced by pulsed modulated millimeter wave heating at cellular level

Figure 3.1: Microscopic view of cultured A375 cell line at (a) low and (b) high density [29].

3.2.2 Exposure protocol For exposure, cells were seeded in 3 wells of 12-well tissue culture plate at a density of 3∙104 cells by well. One well was exposed, one well was used as a negative control, and the third well was used as a positive control, which served as technical verification for detection of heat shock response or apoptosis. Cells were treated with 5 µM of MG132 (Sigma-Aldrich, Saint-Quentin Fallavier, FR) or with 100 µM of Antimycin A (Sigma-Aldrich) to detect HSP27 phosphorylation or Casp-3 activation, respectively. In particular, MG132 was added immediately before the exposure, while Antimycin A was added 24 h before the exposure. As the penetration depth of the MMW is very small (i.e., about 34 mm at 58.4 GHz in culture medium), temperature elevation is very localized, and therefore is expected only in the exposed well, i.e., the one aligned to the open- ended WG axis (Fig. 3.2 - for further details refer to paragraph 2.5.5, Reproducibility). However, to prevent any possible heating of the wells used as controls, cells were plated following the scheme represented in Figure 3.2 (right).

Figure 3.2: Schematic of the exposure system (left), zoom on the exposed well (center), and 12-well tissue culture plate used during exposure of cells (right). CM stands for culture medium.

Data presented in Figure 3.3 show the phosphorylation of HSP27 detected in the negative and positive (e.g., MG132) controls. For each type of control, no significant differences are observed after sham, CW, and PW exposures, i.e., they are independent from the exposure. Data have been normalized to the negative control of the sham exposure. As expected, negative controls show very

151

CHAPTER 3. Cell mortality and heat shock stress induced by pulsed modulated millimeter wave heating at cellular level similar value of phosphorylation of HSP27 as the sham (i.e., ratio is about 1), while cells treated with MG132 demonstrated an increase of the target protein of about 8 times compared to the value of sham, confirming the reliability of our experimental protocol.

Figure 3.3: Phosphorylation of HSP27 analyzed after sham, PW, and CW exposures both in negative and positive (i.e., MG132) controls.

Two days after plating, the medium was replaced by a medium without sodium bicarbonate containing 4.6 mM of Hepes (Thermo Fisher Scientific, Waltham, MA, USA) to maintain constant pH in the non-gassed incubator [30]. Then, the plate with cells was transferred to a standard incubator with exposure facilities. Before exposure, cells were incubated for 1h at 32°C (i.e., average temperature of skin under normal conditions). Heat shock response may appear after a certain delay following the exposure [31], therefore cells were kept at 37 °C for 6 h after exposure. Then, cells were fixed before proceeding to immunochemistry (ICC). Sham exposures were performed under identical experimental conditions, but with the generator switched off. The experimental procedure above described is outlined in Figure 3.4.

152

CHAPTER 3. Cell mortality and heat shock stress induced by pulsed modulated millimeter wave heating at cellular level

Figure 3.4: Time line summarising the exposure protocol. After 48 h from the seeding, DMEM was substituted with DMEM without sodium bicarbonate, just before the exposure of the cells, and MG132 was added in the experiments used to detect HSP27 phosphorylation (**). Antimicyn was added 24 h before the exposure in the experiments used to detect Casp-3 activation (*).

It is worth to note that preliminary studies have shown that 1 h incubation post-exposure is not sufficient to induce differences between the CW and PW exposures (Fig. 3.5), and cellular response observed is less than 1.5 times higher compared to the sham. Therefore, in our experiments we always incubated cells 6 h post-exposure.

Figure 3.5: Phosphorylation of HSP27 analyzed cell-by-cell 1 h and 6 h post exposure to PW and CW exposures with the same average temperature rise (see section 3.3, Electromagnetic and thermal dosimetry). Data are shown as mean values (n = 4) ± SEM normalized to the sham.

3.2.3 Immunocytochemistry protocol ICC is a powerful tool to identify proteins and other macromolecules in tissues and cells [32] by means of antigen-antibody interaction. In the current study, the technique resulted very useful to quantify as well as to analyse the spatial distribution of the specific cellular response of interest. The antigen (i.e., the target molecule such as HSP27 phosphorylated and cleaved Casp-3) is recognized and bound by specific primary antibody. The primary antibody is bound by a secondary antibody labeled with fluorophore to allow the visualization of the target molecule under a 153

CHAPTER 3. Cell mortality and heat shock stress induced by pulsed modulated millimeter wave heating at cellular level fluorescence microscope. The standard ICC protocol involves several steps including cell fixation, permeabilization, blocking, immunolabeling, and counterstaining of the nuclei. Each step requires optimization as experimental variables can significantly impact staining outcome [33]. The ICC and fluorescence analysis protocol used in this study was the same as the one used by Haas et al., [34] to evaluate the potential effects of MMW on PC12 neurite outgrowth and cytoskeleton protein expression, at 60.4 GHz. In this study, the ICC was performed inside the 12- well tissue culture plates. Cells were fixed with 4% paraformaldehyde for 20 min at room temperature, then washed twice with phosphate-buffered saline (PBS), 6 h after exposure. Cellular permeabilization is necessary to facilitate large antibody to cross the cellular membrane. In this study, it was done for 10 min with 0.25% Triton X-100 in PBS. The blocking step minimizes non- specific interactions that result in background and false-positive staining (artifacts). Here, blocking of antibodies unspecific binding was done by incubating cells during 20 min in 1% bovine serum albumin (BSA, MP Biomedicals, Santa Ana, CA), 0.1 % gelatin from cold water fish skin (Sigma- Aldrich, St Louis, MO, USA), and 0.1% Triton X-100 (MP Biomedicals) in PBS. Cells were then incubated overnight, at 4°C, with primary antibodies at 1:200 and 1:500 dilutions, respectively for phosphorylated HSP27 (Phospho-HSP27 (Ser82), ref 2406, Cell Signaling Technology, Danvers, MA) and cleaved Casp-3 (Cleaved Caspase-3 (Asp175), ref 96645, Cell Signaling Technology, Danvers, MA). After three successive washes using PBS supplemented with 0.1% Tween 20 (MP Biomedicals), cells were incubated 1 h with secondary antibodies at 1:1000 dilution and Hoechst 33342 (10 µg/mL, Sigma-Aldrich) for nuclei counterstaining to allow ease identification of individual cells.

Figure 3.6: Schematic of ICC protocol.

154

CHAPTER 3. Cell mortality and heat shock stress induced by pulsed modulated millimeter wave heating at cellular level

3.2.4 Image acquisition and fluorescence analysis Pictures of cells were taken and fluorescence value of each cell was quantified using a Cellomics ArrayScan VTI HCS Reader (Thermo Fisher Scientific) at ImPACcell technological platform (Biosit, University of Rennes 1, Rennes, FR) shown in Figure 3.7. The instrumentation allows high- content fluorescence imaging and quantitative analysis of fixed and live cells. An automatic procedure allowed to take square pictures of 0.25 mm² area, and quantify the fluorescence value of each cell belonging to them. For each well, 121 pictures were taken following a square spiral from the center of the well covering a 30.25 mm² area (Fig. 3.8). This area corresponds, in the case of the exposed well, to the central area of the exposure, i.e., the zone of maximal thermal stress. Picture number 1 of the exposed well was aligned to the open-ended WG axis. As already detailed in the previous chapter, due to the small penetration of the MMW, even a small shift (of the order of few millimeters) between the antenna and the sample may determine large variation of SAR and temperature distributions. This results in different exposure conditions of the sample, thus impacting the cellular response and consequently the interpretation of the biological results. Therefore, in order to assure that all the plates were exposed under the same conditions (i.e., picture number 1 aligned to the WG axis), a plastic frame (Fig. 2.6c) was used to avoid movement of the plate, once the desired position was previously set at the beginning of each set of the experiments (for more details see paragraph 2.5.5, Reproducibility). Moreover, in order to detect the cellular response in the central area of the exposed well, the microscope was aligned to the center of the tissue culture plate. Poor quality pictures were visually screened and deleted.

Figure 3.7: Cellomics ArrayScan VTI HCS Reader [35].

155

CHAPTER 3. Cell mortality and heat shock stress induced by pulsed modulated millimeter wave heating at cellular level

Figure 3.8: Plan of the picture taken within each well of tissue culture plate used for experiments. A central area of 30.25 mm² was covered by taking 121 pictures. Picture number 1 of the exposed well was aligned to the axis of the open-ended rectangular WG. The same plan was used for all the wells allowing comparison of cells in the same positions.

Data of the cellular fluorescence of the desired target molecule within the nucleus and cytoplasm were analysed cell-by-cell (Fig. 3.9) as a function of the distance from the center of the well. Distance equal to 0 mm was assigned to the picture taken in the center of the well (i.e., picture number 1 of Fig. 3.8) aligned to the open-ended WG axis. All the cells included in this picture were considered located at 0 mm. Then, the distance from the center was calculated for all the pictures by using a simple approach (i.e., Pythagoras' theorem) resulting in the plan of the distances of Figure 3.10. Maximal distance of the cells was 3.5 mm with respect to the center. Each picture contained an average number of about 100 cells allowing consistent average of the fluorescence value of the cells at each distance considered in the analysis. It is interesting to note that the area analysed is within the exposure beam which radius, calculated under the same exposure conditions, was found equal to 3.3 mm [36].

156

CHAPTER 3. Cell mortality and heat shock stress induced by pulsed modulated millimeter wave heating at cellular level

(a) (b) Figure 3.9: Fluorescence value analyzed cell-by-cell from 0 to 3.5 mm after (a) CW and (b) sham exposures. Each point represents the value of fluorescence of a single cell at the considered distance.

Figure 3.10: Plan of the distances of the cells within each well with respect to the center (i.e., 0 mm). Zoom on the exposed well is represented.

3.2.5 Statistical analysis Three (n = 3) tissue culture plates were exposed or sham-exposed for each experimental condition. Statistical analyses were performed by using SigmaPlot Statistics or MATLAB tools. As the number of samples was small, the non-parametric Mann-Whitney Rank Sum test was used for 157

CHAPTER 3. Cell mortality and heat shock stress induced by pulsed modulated millimeter wave heating at cellular level the statistical comparisons of the data. For all tests, a p value < 0.05 was considered statistically significant. Results are presented as mean ± SEM (Standard Error of the Mean).

3.3 Electromagnetic and thermal dosimetry

The control of the exposure, both from the electromagnetic and thermal points of view, is outmost importance for a correct interpretation of biological outcomes of the study. Therefore, before exposing cells, careful analysis of the SAR and temperature distributions has been performed by using the methodologies described in Chapter 2. In particular, melanoma cells were exposed in vitro to CW or PW electromagnetic field at 58.4 GHz. The exposure system used in this study has been already detailed in section 2.3 (Experimental set-up) and schematically represented in Figure 3.2. Briefly, cells were exposed from the bottom by an open-ended rectangular WG antenna located 5 mm from the plate inside the incubator, set at 32 °C. As the exposure configuration is the same as the one used to analyse thermal convection, SAR distribution in the exposed culture medium corresponds to the one presented in section 2.6 (Electromagnetic dosimetry). Computed SAR distribution in the cell monolayer was retrieved applying the correction factor extracted from [36]. As detailed in 2.6.1 (Impact of cell monolayer), the SAR distribution in the cell monolayer, under the specific exposure conditions considered in this study, is about 50% lower compared to the one in the cell culture medium (Fig. 3.11).

(a) (b) Figure 3.11: (a) CAD model of the antenna and exposed well were used for computing SAR. (b) Computed SAR in the cell monolayer normalized to the antenna input power of 1W retrieved applying the correction factor extracted from [36]. White ellipses indicate the locations of TC sensors in temperature measurements.

158

CHAPTER 3. Cell mortality and heat shock stress induced by pulsed modulated millimeter wave heating at cellular level

Electromagnetic power dissipated in the cells and culture medium resulted in heating. Local temperature rise in the cell monolayer was measured by using the technique previously detailed (2.5.4, Thermocouple configuration), i.e., using a 75 µm diameter TC. Temperature measurements and cell exposures were performed in separate experiments as we previously demonstrated that the presence of the cells does not perturb the temperature rise in the culture medium (2.8.1, Heating in different media). In this way, possible cellular damage or contamination, as well as local increase of SAR in the cell monolayer due to the presence of the TC (see 2.6, Electromagnetic dosimetry) were avoided. Reproducibility was assured by using the protocol previously described (2.5.5, Reproducibility). Temperature measurements before exposure of cells were repeated at least 4 times if not specified.

3.3.1 1.5 s heat pulses Square-wave amplitude modulation was used to create electromagnetic pulses with the duration of 1.5 s. From now on, heat pulses with duration of 1.5 s will be referred as short pulses. As detailed in the Chapter 2 (2.9.1, Dependence on pulse duration), for this duration of the pulse, the development of convective currents in the exposed liquid has only minor effect. In particular, the temperature rise of the heat pulse is not impacted, while the cooling rate is only slightly affected resulting in decrease of the average temperature for prolonged exposures. The peak power of PW exposure at the WG input was set to 3.7 W to generate the thermal pulses with a peak amplitude of 10°C in the center of the well (Fig. 3.12a). This power corresponds to the SAR of 73.6 kW/kg in the cell monolayer and temperature rise rate of 6.7 °C/s. The period of 20 s was selected to maintain the average PW heating (mean PW in Fig. 3.12c) below 43°C (42.3 ± 0.31 °C), i.e., the thermal threshold above which heat cytotoxicity is markedly higher [37]. The average temperature was calculated by using a moving average filter with 75 s span as already detailed in Chapter 2. The power of CW exposure was adjusted to generate the same heating as the average temperature rise during the PW heating. It was equal to 250 mW corresponding to 4.9 kW/kg in terms of SAR. Exposure duration was set to 90 min corresponding to 270 pulses.

159

CHAPTER 3. Cell mortality and heat shock stress induced by pulsed modulated millimeter wave heating at cellular level

(a) (b)

(c) Figure 3.12: Temperature measured at the bottom of the exposed well. (c) Temperature dynamics at the center of the well for PW and CW exposures. The subplots on the top (a,b) illustrate the zoom for 1–2 min and 89–90 min intervals.

Temperature measurements were performed in different points along the radius of the well, on the E-plane, namely in the center (referred as 0 mm), and at 2.5, 5, 7.5, and 10 mm from the center of the exposed well (see white ellipses in Fig. 3.11b). To assure the reproducibility of the results, the same methodology detailed in 2.5.5 (Reproducibility) was applied. Two different approaches were used. First, two TC were inserted in the exposed well. One was set in the center of the well and used as a control for the correct alignment of the well with the WG axis, and the second was fixed at the desired distance from the center. In these conditions, temperature measurements were repeated at least three times for each position.

160

CHAPTER 3. Cell mortality and heat shock stress induced by pulsed modulated millimeter wave heating at cellular level

The second approach required only one TC. It was attached to the center of the well, and inserted in a specific frame equipped with a micrometer (Fig. 3.13). In this case temperature measurements were performed by moving the tissue culture plate by 1 mm step for each measurement by means of the micrometer. Exposed medium was carefully removed and well rinsed after each exposure. Results obtained with the two methodologies did not revealed significant differences (within the standard deviation of the measurements) guarantying the reproducibility of the temperature measurements presented in this section.

(a) (b) Figure 3.13: Micrometer used for the temperature measurements at different distances.

Results of the measurements performed at different locations along the radius of the well are shown in Figure 3.14a,b. The corresponding SAR for PW and CW exposures was equal to 47.1, 20.6, 8.8, 4.4 kW/kg, and 3.1, 1.4, 0.6, 0.3 kW/kg for 2.5, 5, 7.5, 10 mm, respectively (white ellipses in Fig. 3.11). CW heating and the mean of the PW heating vary equally along the radius of the well. ΔT profile for the peak temperature induced by PW exposure is similar to that of SAR (Fig. 3.14c), maximum relative deviation is of 25% at d = 2.5 mm.

161

CHAPTER 3. Cell mortality and heat shock stress induced by pulsed modulated millimeter wave heating at cellular level

(a)

(b) (c) Figure 3.14: (a) Temperature dynamics at different distances from the center of the well for 90 min of exposure (the locations of TC sensors are schematically illustrated in Fig. 3.11). (b) Corresponding steady- state temperature rise after 90 min of exposure. (c) Normalized SAR along the E-plane and temperature rise ΔT (defined as the difference between the steady-state temperature at considered location and at 10 mm). Error bars indicate the SEM for three independent measurements.

3.3.2 6 s heat pulses In order to study the thermal response of the melanoma cell line as a function of the pulse duration, different durations of the heat pulses were considered. The ad hoc MMW exposure system employed allowed to easily generate a large variety of thermal pulses by adjusting the peak power and the pulse duration of the electromagnetic waveform. It was previously shown that initiation of convection strongly depends on the pulse duration (section 2.9.1, Dependence on pulse duration). Convection largely impacts the heat pulse shape for duration of the pulse > 4.5 s, strongly changing the heat pulse shape, decreasing the peak temperature and increasing the cooling rate. For this reason, preliminary studies were oriented towards the possibility to use high amplitude pulses of duration < 4.5 s, i.e., 3 s, with the same peak power employed for pulses of duration of 1.5 s (3.7 W). Under these conditions, the amplitude of the thermal pulse was about 20 °C. The pulse period was adjusted to 39 s in order to obtain the same average temperature of about 42 °C as the CW heating previously described. Steady state temperature was about 58 °C. Amplitude of the thermal pulse, period, and steady state temperature obtained were 50%, 48%, and about 20% higher, respectively than the correspondent values at pulse duration of 1.5 s (Fig. 3.12). The resulting temperature dynamics during 90 minutes of exposure is represented in Figure 3.15 and corresponded to 138 pulses. In this case, apart from the pulse duration, both electromagnetic (i.e.,

162

CHAPTER 3. Cell mortality and heat shock stress induced by pulsed modulated millimeter wave heating at cellular level pulse period) and thermal parameters (i.e., amplitude of the heat pulse, maximum temperature at steady state, number of the pulses) were varied complicating the interpretation of the results when comparing with the biological outcome of the exposure at shorter duration of the pulse.

Figure 3.15: Temperature dynamics at the center of the well for PW and CW exposures. PW duration was 3 s at the peak power of 3.7 W.

Therefore, despite the interest in understanding the cellular behaviour under 3 s pulses at such high temperature, the decisive idea driving our experiments was to compare thermal pulses of different durations, by varying mainly the electromagnetic parameters of the pulse, including pulse period and peak power, leaving the same temperature dynamics both in terms of peak amplitude of the thermal pulse and maximal temperature rise at steady state. This allowed us to facilitate the interpretation of the results by reducing the number of the variables changed. Pulses with a duration of 6 s were used, as we assumed that pulse duration of 3 s, with the same temperature dynamics as 1.5 pulses, may have not generated notable differences. Duration of the thermal pulse of 6 s results in a great triggering of the convection impacting the shape of the thermal pulse (Fig. 3.16).

163

CHAPTER 3. Cell mortality and heat shock stress induced by pulsed modulated millimeter wave heating at cellular level

Figure 3.16: Variation of the pulse shape as a function of the pulse duration from 1 s to 6 s at the peak power of 3.7 W.

From now on, heat pulses with duration of 6 s will be referred as long pulses. The peak power was reduced to obtain the same peak temperature dynamics in a 6 s pulse as in 1.5 s pulse (Fig. 3.17a), equal to 10 °C. The peak power was equal to 1.6 W (56% lower than the one used in 1.5 s pulse) corresponding to a SAR of 31.8 kW/kg, in the center of the exposure. The period of the pulse was adjusted to obtain the same average temperature rise as for 1.5 s pulses (Fig. 3.14b). It was equal to 27.9 s (28% less of the one used for short pulses) corresponding in total to 194 pulses (28% less of the number of pulses for the duration of 1.5 s) for 90 min of exposure. Under these conditions, the temperature rise rate was 1.7 °C/s (75% less of the short pulses). Note that in order to obtain the same average temperature dynamics for both the pulse durations considered in the study, the average power used for the longest pulse was higher by about 20% compared to the one used during exposure to short pulses (Table 3.1). This allowed to compensate for the effects of convection, which tends to decrease the peak amplitude and increase the cooling dynamic of the longer thermal pulse, resulting in decrease of the average temperature for prolonged exposure. Parameters of the electromagnetic pulses used in the experiments are summarized in the Table 3.1.

164

CHAPTER 3. Cell mortality and heat shock stress induced by pulsed modulated millimeter wave heating at cellular level

(a) (b) Figure 3.17: Comparison of measured at the center of the well temperature dynamics for 1.5 s and 6 s pulses with the same peak and average temperatures. (a) Single pulse heat profiles. (b) Temperature dynamics for 90 min of exposure.

Table 3.1. Parameters of the electromagnetic pulses for the total exposure duration of 90 min

Pulse Temperature Peak Average Pulse Period Duty cycle Number of duration rise rate power power energy (s) (%) pulses (s) (°C/s) (W) (W) (J) 1.5 6.7 20 7.5 270 3.7 0.275 5.55 6 1.7 27.9 21.5 194 1.6 0.334 10.2

Temperature measurements were also performed at different points within the analysed area, namely at 1, 2, and 3 mm from the center of the well. Results show that the maximum temperature reached during exposure to short pulses is 49 ± 0.1 °C, 48.3 ± 0.3 °C, and 46 ± 0.23 °C, respectively at 1, 2, and 3 mm from the center. Temperature measurements obtained for longer pulses evidenced that convection triggered in the cell culture medium slightly decreases the peak temperature compared to shorter pulses (Fig. 3.18). Maximum relative deviation is of 27% at d = 2 mm.

165

CHAPTER 3. Cell mortality and heat shock stress induced by pulsed modulated millimeter wave heating at cellular level

Figure 3.18: Normalized temperature rise ΔT (defined as the difference between the steady-state temperature at considered location and at 3 mm) of the maximum temperature reached during the PW exposures. Error bars indicate the SEM for three independent measurements.

Minimum temperature, for both the pulse durations considered, varies of 0.5 °C between 0 and 3 mm from the center. Differences observed are within the standard deviation of the measurements (data not shown). The average temperature obtained for short pulses, as expected from previous measurements (Fig. 3.14d) varies proportionally to CW heating for all the distances (Fig. 3.19). Maximum relative deviation is of 27% at d = 3 mm (0.3 ± 0.1 °C). In the case of long thermal pulses, the average temperature recorded in the center of the exposure is about 0.3 ± 0.15 °C higher than the corresponding temperature of short pulses (Fig. 3.19). We assume that this slight increase is due to the slightly higher average power used to compensate for the effect of the thermal convection to maintain the same temperature dynamics. This difference decreases at points further from the antenna. Data are shown as a linear regression of the normalized ΔT defined as the difference between the steady-state temperature at considered location and at 3 mm. For all the exposures, the average temperature difference between the center and 3 mm from the center of the well is within 1 °C.

166

CHAPTER 3. Cell mortality and heat shock stress induced by pulsed modulated millimeter wave heating at cellular level

Figure 3.19: Linear regression of the normalized ΔT (defined as the difference between the steady-state temperature at considered location and at 3 mm) of the average temperature reached during the PW and CW exposures. Error bars indicate the SEM for three independent measurements.

3.3.3 Thermal dose Precise measurements of the temperature reached during the exposure allowed us to obtain information about the real time thermal dose accumulated by the cells during the exposure. In order to compare the different exposure protocols employed in this study we used the degree x minutes (dm43) and the cumulative equivalent minutes at 43°C (CEM43°C) [37]. The degree x minutes has been calculated as:

(3.1)

where n represents the number of intervals in which the duration of the exposure has been divided,

is the average temperature in the n-th time interval, 43°C is the reference temperature [37], and is the starting temperature (i.e., 32°C) before exposure. In our calculation, in order to take into account the fast temperature variations during the pulse exposure, the averaging interval was fixed to 0.3 s. The cumulative equivalent minutes at 43 °C (CEM43°C) has been calculated as:

167

CHAPTER 3. Cell mortality and heat shock stress induced by pulsed modulated millimeter wave heating at cellular level

(3.2)

where n represents the number of intervals in which the duration of the exposure has been divided,

is the i-th interval over temperature is averaged (in minutes), is the average temperature during the time interval , 43°C is the reference temperature [37], and R is the factor to compensate for a 1°C temperature change and it is related to the temperature dependence on the rate of cell death [37]. The parameter R used in this study has been taken from [38], derived from skin human cells in vitro:

(3.3)

In order to take into account the fast temperature variations during the pulse exposure, the averaging interval was fixed to 0.3 s.

3.4 Heat shock response

It has been demonstrated by previous studies that in case of heat shock, cells overexpress chaperones and heat shock proteins (HSP) that protect cellular proteins from misfolding and aggregation [22]. HSP have been identified as key determinants of cell survival because they may modulate apoptosis by directly interacting with the components of the apoptotic machinery [39]. In particular, physiological stimuli (such as redox signaling, cytokines, and growth factors) and various forms of stress (e.g., heat) dramatically increase the phosphorylation of some small HSP such as HSP27 [40]. Protein phosphorylation is the most widespread post-translational modification in eukaryotic cells, and it is involved in all fundamental cellular processes including signal transduction, cell cycle, growth, apoptosis, membrane transport, protein degradation, enzyme regulations, and protein-protein interactions [41]–[43]. The process is regulated by a family of enzymes named protein kinases [44] and consists in the addition of a phosphate group (PO4) to the polar group R of various amino acids. In eukaryotic cells, it typically occurs at the side chains of three amino acids, namely serine, threonine, and tyrosine [45], [46]. The addition of the phosphate group modifies the protein from hydrophobic apolar to hydrophilic polar inducing conformational changes [41] that may affect their functions when interacting with other molecules. The phosphorylation is reversed

168

CHAPTER 3. Cell mortality and heat shock stress induced by pulsed modulated millimeter wave heating at cellular level by another family of enzymes, called protein phosphatases. They remove the phosphate group from phosphoproteins by hydrolysing phosphoric acid monoesters into a phosphate group and a molecule with a free hydroxyl group [47]. Up to 30% of all human proteins may be modified by kinase activity [44], [46]. In particular, phosphorylation plays a crucial role in the cellular functions of small HSP, and particularly in HSP27 [48]. The latter represents the most abundant small HSP in humans and is expressed under basal conditions levels in cells and tissues [49]. HSP27 exhibits a monomeric molecular mass of 15–42 kDa, and can form oligomeric complexes of up to 80 kDa [39]. This protein is up-regulated by heat shock [50], oxidative stress [51], during aging [52], in cancers [53], protein deposition diseases [54], and ischemic neuronal injury [55]. It is phosphorylated by MAPKAP (mitogen- activated protein kinase-activated protein) kinase 2/3 via the activation of the p38 MAPK pathway at multiple serine residues [56] in response to various cellular stresses including heat shock, mitogens, cytokines, lipopolysaccharides, phorbol esters, thrombin, serotonin, angiotensin, vasopressin, endothelin, carbachol, ceramide, glucose, metals, vitamin D3, and retinoic acid [48]. Following phosphorylation, HSP27 reorganizes itself into smaller oligomers, often dimers and tetramers, and can reprogram its interaction with specific partner/client polypeptides [39] (Fig. 3.20). In particular, this dynamic change allows the HSP27 to trap and store misfolded polypeptides, and then refold them through cooperation with the well characterized ATP-dependent “foldase” chaperone machinery refolding activity [57].

Figure 3.20: Structural organization of Hsp27 upon reversible phosphorylation (adapted from [42]).

Several cellular functions are associated with HSP27 phosphorylation including actin filament dynamics, anti-proliferation, intracellular virus transport, cellular differentiation, and cell survival [48]. The anti-apoptotic function is the most recognized [48]. It is exerted by HSP27 phosphorylated by interacting with both mitochondrial dependent and independent pathways of apoptosis (3.5.1, Intrinsic and extrinsic apoptotic pathways). For example, it has been demonstrated that phosphorylated dimers of HSP27 interact with Daxx, a mediator of Fas-induced apoptosis,

169

CHAPTER 3. Cell mortality and heat shock stress induced by pulsed modulated millimeter wave heating at cellular level preventing the interaction of Daxx with both Ask1(Apoptosis signal-regulating Kinase 1 [58]) and Fas and blocking Daxx-mediated apoptosis [59]. HSP27 may also interact with Bax and cytochrome c, preventing mitochondrial dependent apoptosis [60]–[62]. Moreover, it has been also demonstrated that phosphorylation of HSP may inhibit the Caspase-dependent apoptosis by repressing the Casp-3 activation [31]. However, several in vitro studies also demonstrated that large oligomeric structures of the molecules facilitate the chaperone activity, i.e., misfolded proteins are easily stored and refolded by the HSP, while protein redistribution in small oligomers due to excessive phosphorylation, seem to negatively regulate the chaperone-like activity of HSP27 leading to a significant decrease in their protective function consisting in preventing thermal aggregation of the proteins [63]–[65].

3.4.1 Dose-response curves for continuous wave heating In this study, phosphorylation of HSP27 was used as a marker of heat-induced cellular stress. Preliminary studies were performed with the attempt to investigate the cellular response of the melanoma cell line to continuous heat in the 37–47 °C temperature interval after exposure of 90 minutes. Desired temperature dynamics were obtained by varying the input power of the open- ended WG between 120 and 430 mW corresponding to SAR values between 3.8 and 16.4 kW/kg (Fig. 3.21). Temperature curves refer to point measurements in the center of the exposed well.

Figure 3.21: Temperature dynamics during 90 min of exposure to CW in the 3.8–16.4 kW/kg SAR interval. Results are given as mean (± 0.3 °C) of all the measurements for each condition.

170

CHAPTER 3. Cell mortality and heat shock stress induced by pulsed modulated millimeter wave heating at cellular level

Heat shock response was quantified 6 h after exposure. The phosphorylation of HSP27 has been analysed as a function of the distance from the center of the exposure. Data are presented as fold of induction of the sham. They show linear increase of the protein phosphorylation (Fig. 3.22a) as the thermal dose (Fig. 3.22b), i.e., the temperature, increases. Results demonstrated that in the central area of the exposure (0 < d < 1.8 mm) a clear heat shock response appears at temperatures > 41 °C.

(a)

(b) Figure 3.22: (a) Phosphorylation of HSP27 as a function of the temperature in the 37–47 °C temperature range, for the A375 melanoma cell line. (b) CEM43°C calculated by using the temperature dynamics curves in the 37–47 °C temperature interval (Fig. 3.21). 171

CHAPTER 3. Cell mortality and heat shock stress induced by pulsed modulated millimeter wave heating at cellular level

Moreover, the heat shock response of the A375 melanoma cell line has been analysed between 0 and 3.5 mm with higher spatial resolution (Fig. 3.23). Data showed that phosphorylation of HSP27 was independent from the distance up to 40 °C. When CW heating provided was 41 °C, a slight increase (maximum value of about 1.2-fold of the sham) of the cellular response is observed in the close proximity of the WG axis. At CW exposure of 42 °C, heat shock response becomes clearly higher compared to the one of the sham in the center of the well, and it rapidly decreases while moving further from the WG. At higher temperatures, HSP27 phosphorylation remains higher than sham within the whole area analyzed, with values of induction up to 40-fold the sham (i.e., at 46°C and 47 °C).

Sham 37 °C 38 °C

39 °C 40 °C 41 °C

172

CHAPTER 3. Cell mortality and heat shock stress induced by pulsed modulated millimeter wave heating at cellular level

42 °C 43 °C 44 °C

45 °C 46 °C 47 °C Figure 3.23: Phosphorylation of HSP27 analyzed cell-by-cell 6 h post exposure to CW in the 37–47 °C temperature interval. Different scales are used for the vertical axis.

3.4.2 Pulsed vs continuous heating The knowledge of the heat shock response of cells represents a fundamental step towards the planning of a specific thermal treatment of cancer. It was shown by previous studies that the response to treatment may vary considerably among different cell lines due to intrinsic differences which affect the heat sensitivity of the tumor cells [66]. Therefore, the first part of this study has been dedicated to the understanding of the thermal stress of the A375 melanoma cell line following the exposure to CW and PW regimes, with the same average temperature dynamics. As we previously demonstrated that the heat shock response mediated by the phosphorylation of the HSP27 in A375 melanoma cell line is triggered at temperatures higher than 41 °C (Fig. 3.23), the reference temperature used in this study for the PW average as well as for the CW heating, was about 42 °C (Fig. 3.17b). Moreover, the latter represents typical temperature employed in moderate hyperthermic oncological therapies alone [67] or in combination with other treatments [68]. Figure 3.24 represents the spatial distribution of the fluorescence intensity of cells following PW, CW, and sham exposures in the central area of the well (Fig. 3.10). It qualitatively shows that PW heating induces stronger phosphorylation of HSP27 compared to CW heating. In particular, thermal 173

CHAPTER 3. Cell mortality and heat shock stress induced by pulsed modulated millimeter wave heating at cellular level response produced by short pulses with duration of 1.5 s appears stronger and focused towards the location of the antenna (i.e., the center of the well), while the one induced by long pulses of 6 s seems to spread almost over the whole region of interest.

Figure 3.24: Spatial distribution of the normalized intensity of the HSP27 phosphorylation for a) PW (1.5 s), b) PW (6 s), c) CW, and d) sham exposures.

In the area up to 1.8 mm from the center of the well, pulses with duration of 1.5 s resulted in values of phosphorylation of HSP27 significantly higher than the ones induced by PW with duration of 6 s and CW heating (Fig. 3.25a). Within this area, the phosphorylation of HSP27 was 10.1, 7.9, and 6.4-fold induction in respect to sham for PW with duration of 1.5 s, PW with duration of 6 s, and CW, respectively. For distances exceeding 1.8 mm (SAR, ΔT [PW max], and ΔT [CW] values are below 77%, 86%, and 96% in respect to the peak on the axis, respectively) it decreased to 4.3, 5.3, and 2.2-fold induction, for short pulses, long pulses, and CW heating, respectively. Cellular response induced by both durations of the heat pulses remains higher compared to CW heating within the whole area analyzed. However, differences were observed between short and long pulses. In the case of short pulses, heat shock response is significantly higher by a factor of 1.6 and 1.9 than CW, in the areas between 0 and 1.8 mm, and 1.8 and 3.5 mm, respectively. In the case of long pulses, heat shock response is 1.2 times higher than CW in the 0–1.8 mm area, and difference is not statistically significant. In the region between 1.8 and 3.5 mm, this difference increases (i.e., 2.6 times higher) acquiring statistically significance. Moreover, in the same region, 174

CHAPTER 3. Cell mortality and heat shock stress induced by pulsed modulated millimeter wave heating at cellular level thermal response induced by long pulses became 1.4 times higher than short pulses, as opposite from the behaviour observed in the region up to 1.8 mm (short pulses-induced phosphorylation is 1.3 times higher than the one induced by long pulses).

(a)

(b) Figure 3.25: Phosphorylation of HSP27 after PW and CW exposures with the same average temperature rise. (a) Phosphorylation of HSP27 analyzed cell-by-cell 6 h post exposure shown as mean values (n=3) ± SEM normalized to the sham. The data are averaged over the areas around the center of the well (i.e., 0– 1.8 mm denotes the data averaged over the area with the radius of 1.8 mm, and 1.8–3.5 denotes the data averaged over 1.8 mm to 3.5 mm from the center). Asterisk (*) indicates statistical significance at p < 0.05 compared to the CW, while double asterisk (**) indicates statistical significance at p < 0.05 compared to the PW with duration of 6 s. (b) The same data shown for the averaging with higher spatial resolution.

The analysis of the phosphorylation of HSP27 with distance revealed important features of the cellular response to heat pulses at different durations compared to CW heating. The level of phosphorylation of HSP27 induced by the CW exposure presents steeper reduction of its value with 175

CHAPTER 3. Cell mortality and heat shock stress induced by pulsed modulated millimeter wave heating at cellular level the distance (i.e., temperature) compared to the PW heating, for both the durations of pulses. In the area from 0 to 3.5 mm, the heat shock response decays by a factor of 4.2, 2.7, and 1.4, following CW, short, and long PW-induced heating, respectively (Fig. 3.25 b). In the same area, the SAR and thermal pulse amplitude drop were about 50%, while the decrease of the CW heating was around the 80% (Fig. 3.14c). As shown by preliminary results, the phosphorylation of HSP27, for the specific cell line used in the study, is triggered at continuous temperature > 41°C in the center of the well, decreasing rapidly in cells located at distances > 1 mm from the center of the exposure. It is interesting to note that the phosphorylation of HSP27 presents a plateau between 0 and 1.8 mm from the center of the well after the exposure of cells to PW regime (cellular response decreased only by 1.1 and 0.95 times for short and long pulses, respectively), which is absent in the cells exposed to CW regime. We assumed that this plateau probably corresponds to the phosphorylation of all HSP27 proteins present in cells, as confirmed by the activation of the Casp-3 (see 3.5, Cell mortality). HSP27 phosphorylation of the sham exposed cells did not show any dependence on the distance confirming the reliability of the experimental protocol. An example of the cells located in the central point of the exposure, aligned to the WG axis (i.e., picture number 1 of Fig. 3.8) is represented in Figure 3.26. The blue represents the cellular nucleus, while the green fluorescence is the dye associated to the target protein, i.e., HSP27 phosphorylated.

PW (1.5 s) PW (6 s)

176

CHAPTER 3. Cell mortality and heat shock stress induced by pulsed modulated millimeter wave heating at cellular level

CW Sham Figure 3.26: Fluorescence microscopy pictures of A375 cells for PW (1.5 s), PW (6 s), CW, and sham. Each picture is taken in the center of the exposure (i.e., picture number 1 of Fig. 3.8). Blue represents the nucleus, while the green colour associated with cells represents the fluorescence due phosphorylated HSP27.

Results of this study demonstrated that long duration of the heat pulses results in a larger extension of the heat shock response compared to short pulses with similar temperature dynamics. We assume that among other parameters (i.e., peak power, duty cycle, etc.), the duration of the thermal pulse plays a fundamental role in determining the spatial extension of the cellular response in terms of thermal stress. As suggested in [69], if the duration of the pulse is shorter than the thermal diffusion time, the distribution of the thermal energy is confined within the irradiated zone. Pulses longer than the thermal relaxation time lead to diffusion of heat outside the affected zone causing, in the practical clinic, the damage of the neighbouring healthy tissue. The thermal diffusion time is defined as:

(3.4)

where (mm) is the penetration depth of the incident wave and (m2/s) is the thermal diffusivity of the culture medium. The latter may be modelled as water as differences, in terms of electromagnetic and thermal properties, are within the 3%. The critical time calculated in this study is equal to 0.12 s, obtained for equal to 0.26 mm and equal to 0.143 mm²/s. Thermal diffusion time is much lower than the duration of the pulses employed in this study. However, duration of long pulses is 50-fold and duration of short pulses is 12.5-fold greater than the thermal diffusion

177

CHAPTER 3. Cell mortality and heat shock stress induced by pulsed modulated millimeter wave heating at cellular level time. Heat pulses with short duration deliver high peak power to the target location restricting the heat affected zone to a localized region as opposite to the long pulse duration where heat is more widespread. Moreover, the higher heat shock response observed after 6 h of exposure to long pulses, in cells located at points further from the WG (d > 1.8 mm) may be associated to thermodynamic phenomena occurring in the exposed liquid. Indeed, we cannot exclude that thermal convection, facilitated by the long duration of the thermal pulse may be also responsible of the higher heat diffusion within the liquid. In addition, movement of the liquid bulk, due to the development of convective currents, may modify the conformation of the proteins of the serum, possibly determining different cellular response compared to the cells exposed to shorter pulses, with the same temperature dynamics.

3.5 Cellular apoptosis

HSP expression in cells may correlate with healing or with tissue damage. Therefore, the complete comprehension of the cellular response to heat requires also the evaluation of the cellular mortality. In this study, with the attempt to correlate the level of the heat stress induced by CW and PW-modulated regimes on melanoma cells with the effective induced damage, cellular apoptosis (i.e., cellular programmed cell death) was also quantified under the same experimental conditions as for the detection of the heat shock response. Several tests were preliminarily performed to assess the best compatible assay with our cell line and exposure protocol. It is worthwhile to note, that due to technical constraints, live cells could not be directly analysed, meaning that all the experimental protocols used in this thesis required cellular fixation. Two types of tests were performed on cells including i) cell viability assays based on the loss of membrane integrity (i.e., SYTOX® green [70] and LIVE/DEAD® [71] tests), and ii) ICC protocols aiming to detect some of the principal components activated by the apoptotic machinery (i.e., Bax and Casp-3).

SYTOX® green Loss of membrane integrity is an indicator of cell death. SYTOX® Green [70] nucleic acid stain

(Thermofisher Scientific, Waltham, MA, USA) is a high-affinity nucleic acid stain that easily penetrates cells with compromised plasma membranes and will not cross the membranes of live cells. Cells that exclude a dead cell dye are considered viable, while cells with compromised

178

CHAPTER 3. Cell mortality and heat shock stress induced by pulsed modulated millimeter wave heating at cellular level membranes allow the dye inside the cell to stain an internal component, thus identifying the dead cells. This technique was not compatible with cellular fixation. We assumed that the process of fixation, indeed, created holes on the membrane which facilitated the entering of the dye even in the case of live cells.

LIVE / DEATH®

LIVE/ DEAD® assay (Thermofisher Scientific) allows to differentially label live and dead cells with fluorescent dyes with a one-step live dead assay protocol. The assay staining solution is a mixture of two fluorescent dyes that differentially label live and dead cells, based on the integrity of their membrane. As for the SYTOX® GREEN test, preliminary results demonstrated the incompatibility of the assay with our exposure protocol.

BAX and Caspase-3 Several in vitro and in vivo studies identified apoptosis as the key event responsible for induction of the cell death in response to thermal stress [72], [73]. As the previous tests, based on the membrane integrity did not provide informative results, new protocols were assessed based on ICC for the detection of some of the components involved in the apoptotic machinery, including Bax [74] and Casp-3 [75] (3.5.1, Theoretical basis of cellular apoptosis). It was found that the best way to assess the relative apoptotic activity in high-content fluorescence microscopy analysis is to calculate the percentage of cells above a certain threshold of activate Casp-3 labeling [76]. Therefore, in this study we quantified Casp-3 activation and expressed our data as a percentage of positive cells which represented the percentage of apoptotic cells.

3.5.1 Theoretical basis of cellular apoptosis Apoptosis or programmed cell death is a regulator of physiological growth control and tissue homeostasis [77]. It is characterized by typical morphological and biochemical hallmarks, including cell shrinkage, hypercondensation of chromatin, cleavage of chromosomes into nucleosomal fragments, membrane blebbing, and packaging of cellular content into membrane-enclosed vesicles or “apoptotic bodies” which are phagocytosed by surrounding cells [78], [79]. Cellular apoptosis is a highly regulated process primarily orchestrated by the activation of a cascade of an evolutionary cysteine-dependent aspartate-driven proteases (Caspases), whose name is due to their ability to cleave after aspartic acid [80] promoting cellular death [17]. They are

179

CHAPTER 3. Cell mortality and heat shock stress induced by pulsed modulated millimeter wave heating at cellular level named as Caspase plus a number assigned based on the date of publication (i.e., Casp-2, Casp-8, etc.) [80]. Caspases pre-exist in cells as catalytically inactive zymogens (pro-Caspases) precursors, activated in response to a variety of cell death stimuli. They are divided into two groups, namely initiator and effector Caspases, both composed of three different subunits including a prodomain and two catalytic subunits (Fig. 3.26) [81]. Initiator Caspases have a large prodomain that interacts with a specific adaptor protein required for dimerization-induced activation. The dimerization of the initiator Caspases takes place within a specific protein complex (e.g., DISC for Casp-8, apoptosome for Casp-9, etc. [81]). Once activated, the initiator Caspase cleaves effector Caspase separating the large and small subunits [79]. Following cleavage, effector Caspases act directly on specific cellular substrates to dismantle the cells as detailed in [82]. Initiator Caspases auto-proteolytically cleave whereas executioner Caspases are cleaved by initiator Caspases.

Figure 3.27: Structure of initiator and effector Caspases including a prodomain, large and small subunits. Initiator Caspases have a long prodomain which plays a crucial role for proximity-induced activation mediated by the interaction with adaptor proteins. Two distinct, but structurally related, propeptides have been identified: the Caspase Recruitment Domain (CARD) and the Death Effector Domain (DED). These domains typically facilitate interaction with proteins that contain the same motifs [72]. The effector Caspases (i.e., Casp-3,-6, and 7) have a short prodomain, while the initiators have a long prodomain. Casp- 1,-4,-5, and -12 are implicated in inflammation, whereas Casp -8/-10 and Casp -2/-9 initiate the extrinsic and intrinsic apoptosis pathways, respectively [81].

Initiation of apoptosis occurs through either intrinsic or extrinsic pathways depending on the cellular stimulus triggering the initiation of the process. They are schematically represented in Figure 3.28. All pathways lead to the activation of the major effector Caspases, including Casp-3, - 6, and -7 which coordinate the demolition of key cellular structures and organelles, i.e., by providing DNA condensation and degradation [82], [83] or the cytoskeletal reorganization and disintegration of the cell into apoptotic bodies [84].

180

CHAPTER 3. Cell mortality and heat shock stress induced by pulsed modulated millimeter wave heating at cellular level

Extrinsic pathway is initiated by extracellular signals at the level of the plasma membrane involving the binding of extracellular death ligands (i.e., FasL or tumour necrosis factor-α (TNFα)) to its cell surface transmembrane "death receptor" (i.e., Fas) causing oligomerization of the receptor [77]. Death receptors are members of the tumor necrosis factor (TNF) receptor gene superfamily [85] sharing similar, cysteine-rich extracellular domains. In addition, death receptors are defined by a cytoplasmic domain of about 80 amino acids called ‘death domain’, which plays a crucial role in transmitting the death signal from the cell’s surface to intracellular signaling pathways. The best- characterized death receptors include CD95 (APO-1/Fas), TNF receptor 1 (TNFRI), TNF-related apoptosis-inducing ligand-receptor 1 (TRAIL-R1) and TRAIL-R2. The corresponding ligands of the TNF superfamily comprise death receptor ligands such as CD95 ligand (CD95L), TNFα, lymphotoxin-α, TRAIL, and TWEAK [77]. Binding of ligands (such as CD95L) to death receptors (such as CD95) results in the clustering of the receptors’ death domains and recruitment of adaptor molecules such as Fas-associated death domain (FADD) through homophilic interaction [86] mediated by the death domain [87]. FADD, in turn recruits Casp-8 to the activated CD95 receptor to form the CD95 death-inducing signaling complex (DISC) [88].This, in turn, promotes clustering of proteins that bind to the intracellular domain of the receptor (e.g., FADD, or Fas-associated death domain-containing protein), which then binds to the prodomain of initiator Caspases (e.g., Casp-8 or -10) to promote their dimerization and activation. Casp-8 and -10 are activated within the DISC through induced proximity dimerization [89], and as initiator Caspases, they can successively directly cleave and activate effector Caspases, such as Casp-3. In some cell types, this pathway is sufficient to cause cell death (Type I cells). However, in other cells, (Type II cells), Casp-8 may trigger the intrinsic apoptosis pathway through Bid cleavage [90], [91]. Intrinsic pathway is activated by a variety of stress conditions such as UV irradiation, growth factor deprivation, abnormal mitosis, anti-cancer drug, oxidative stress, and starvation. The process is triggered at the level of the mitochondria. In this case the activation of the Caspase cascade is typically linked to permeabilization of the mitochondrial outer membrane (MOM) regulated by the Bcl-2 family whose members fall into three subclasses: 1) the pro-apoptotic BH3-only proteins (i.e., Bid, Bim, Puma, Noxa, Bad, Bmf, Hrk, and Bi); 2) the prosurvival Bcl-2-like proteins; 3) the pore- forming Bax and Bak proteins [92]. For instance, upon apoptotic stimuli, Bax and Bak are activated and oligomerize at MOM to mediate its permeabilization [74], leading to the release of a set of proteins including cytochrome c, Smac/DIABLO, Omi/HtrA2, Apoptosis Inducing Factor (AIF), and endonuclease G [93]. Once in the cytosol, these apoptogenic proteins trigger the execution of cell death by promoting Caspase activation, or by acting as Caspase-independent death effectors [94]. For example, the release of cytochrome c triggers Caspase activation through formation of the 181

CHAPTER 3. Cell mortality and heat shock stress induced by pulsed modulated millimeter wave heating at cellular level cytochrome c/Apaf-1/pro-Casp-9 apoptosome complex [95]. In particular, the formation of apoptosome is due to the binding of cytochrome c to the C-terminal region of Apaf-1 (apoptotic protease activating factor-1). This facilitates the association of dATP with Apaf-1 inducing conformational changes in the molecule that recruits and activates Casp-9. The latter activates downstream1 Caspases with a short prodomain, such as Casp-3, that successively cleaves key substrates in the cell which either are activated to promote different aspects of cell killing or are inactivated to facilitate cell killing because these substrates are essential for the maintenance and survival of the cell [96]. Other protein released from mitochondria, such as Smac/DIABLO and Omi/HtrA2, facilitate Caspase activation through neutralizing endogenous inhibitors of Caspases, the inhibitor of apoptosis proteins (IAPs) [77].

Figure 3.28: Outline of intrinsic and extrinsic apoptosis pathways ending with the cleavage of the Casp-3.

3.5.2 Pulsed vs continuous heating Results of the apoptotic response of the A375 melanoma cells mediated by cleaved Casp-3, following the exposure to CW and PW- induced heating, with the same average temperature rise, are reported and discussed in this section. Comparison with the heat shock response is presented

1 In molecular biology the definition "acting downstream" is used with the meaning of temporally after. 182

CHAPTER 3. Cell mortality and heat shock stress induced by pulsed modulated millimeter wave heating at cellular level with the attempt to provide complete overview about the cellular response of the cell line of interest under the exposure conditions of the study. Figure 3.29 illustrates the distribution of apoptotic cells according to the Casp-3 activation detection following PW, CW, and sham exposures. It qualitatively shows that only thermal pulses with duration of 1.5 s are able to induce stronger apoptosis, while thermal pulses with duration of 6 s as well as CW heating with the same average temperature rise (Fig. 3.17b) do not induce any noticeable effect.

Figure 3.29: Spatial distribution of apoptotic cells for a) PW (1.5 s), b) PW (6s), c) CW, and d) sham exposures.

In the central area of the well, i.e., 0 < d < 1.8 mm, apoptosis triggering was almost 5 and 2 times greater for PW with duration of 1.5 s compared to long heat pulses of 6 s and CW heating (Fig. 3.30a). This effect rapidly disappears at d > 1.8 mm: apoptotic cells ratio decreased to values < 1.5%, and the difference between PW and CW inductions becomes statistically non-significant for all the exposures considered. The very small baseline cellular apoptosis after sham exposure (< 1%), not dependent on the distance, suggests reliability of the experimental protocol.

183

CHAPTER 3. Cell mortality and heat shock stress induced by pulsed modulated millimeter wave heating at cellular level

(a)

(b) Figure 3.30: Percentage of apoptotic cells after PW and CW exposure with the same average temperature rise. (a) Apoptotic response analyzed cell-by-cell 6 h post exposure shown as mean values (n=3) ± SEM normalized to the sham. The data are averaged over the areas around the center of the well (i.e., 0–1.8 mm denotes the data averaged over the area with the radius of 1.8 mm, and 1.8–3.5 denotes the data averaged over 1.8 mm to 3.5 mm from the center). Asterisk (*) indicates statistical significance at p < 0.05. (b) The same data shown for the averaging with higher spatial resolution.

Interestingly, for heat pulses with duration of 1.5 s, the percentage of apoptotic cells was reduced faster with distance from the WG axis (Fig. 3.30b) compared to the SAR and the thermal pulse amplitude drop (Fig. 2.14d) and is not linear. The percentage of apoptotic cells due to PW exposure was reduced by almost 7 times from 0 to 3.5 mm, while the SAR and the thermal pulse amplitude drop were only around 50%. Very similar behaviour was shown by cells exposed to heat pulses of duration of 6 s. Their apoptotic response was reduced by a factor of 5.5 from 0 to 3.5 mm, resulting steeper than the SAR and thermal pulse amplitude decay. This non-linear dependence of apoptotic cellular response on the thermal pulse amplitude may be explained by sensitivity of the melanoma 184

CHAPTER 3. Cell mortality and heat shock stress induced by pulsed modulated millimeter wave heating at cellular level cells to the properties of thermal pulses (e.g., duration, amplitude, period, temperature rise rate and fast on/off temperature change) or, most probably, by the existence of threshold in the capacity of cells to respond to thermal stress, such as limits in HSP regulation. In particular, we have previously shown that within the 0–1.8 mm region, the phosphorylation of HSP27 due to PW induction, presents a plateau, which corresponds to a drastic increase of apoptosis (Fig. 3.30b), consistent with the fact that refolding system is saturated. Indeed, although HSP permit to protect cells against lethal thermal injury, their induction and related thermotolerance indicate that significant injury has already occurred at the cellular level [97]. In this study, the highest level of phosphorylation of HSP27 induced by short heat pulses (Fig. 3.24) is related to the strongest cellular death (Fig. 3.30) compared to both long heat pulses and CW heating. Our data suggest that when cells are exposed to temperature pulses reaching a maximum peak temperature higher than 48 °C, under certain conditions of the thermal pulse (i.e., pulse duration, temperature rise rate, duty cycle, etc.) they are not able to cope with the thermal stress fully activating repair processes, resulting in initiation of an apoptotic pathway as visualized by the cleavage of Casp-3. The reduction with distance (i.e., temperature) of Casp-3 activated by the CW exposure was slower in comparison with the correspondent HSP27 phosphorylation. Apoptotic cellular response after CW exposure was reduced by about 2.6 times between 0 and 3.5 mm, showing a variation nearly proportional to SAR and temperature drop. The purely thermal mechanism implies the proportional dependence of the effect on absorbed energy. Results suggest the thermal effect on Casp-3 activation is different between CW and PW. In contrast to the PW heating, the CW heating provided in this study corresponds to a mild heat shock (between 42.3 °C and 41.6 °C in the whole area of interest) treatment which generates only a thermotolerance response [97] as visualized by HSP27 phosphorylation (Fig. 3.24) which inhibits the activation of the Casp-3 (Fig. 3.30). This is in agreements with numerous previous studies [98]–[102] demonstrating that exposure of cells to non- lethal temperatures (i.e., < 43°C) may initiate the cellular self-repair mechanism mediated by the induction of molecular chaperones (i.e., HSP) preventing cells from protein denaturation and apoptosis [103]. In particular, it has been demonstrated that the induction of HSP27 may interact with the apoptotic machinery, exercising anti-apoptotic activity, by following different paths. For example, several studies evidenced that HSP27 interferes specifically with the mitochondrial (i.e., intrinsic) pathway of Caspase-dependent cell death [31], [60]. Pandey et al. [31], showed that HSP27 functions as an intracellular inhibitor of Casp-3 activation acting downstream to mitochondrial release of cytochrome c, Apaf-1, and Casp-9 activation (Fig. 3.31). Other studies, performed by the group of Garrido et al., on U937 human leukemic cells [60], [104], demonstrated that HSP27 185

CHAPTER 3. Cell mortality and heat shock stress induced by pulsed modulated millimeter wave heating at cellular level overexpression binds to cytochrome c, preventing the formation of the apoptosome (by inhibiting activation of pro-Casp-9) thus inhibiting the activation of the Caspase cascade ending with the cleavage of Casp-3. Moreover, it has been demonstrated that phosphorylated HSP27 may indirectly inhibits the apoptotic events by triggering the activity of Protein Kinase B (PKB) [105], generated in response of growth factors and cellular stress. PKB, in turn, prevents the release of cytochrome c from the mitochondria [106], blocking the whole chain of events leading to cellular death. Phosphorylated dimers of HSP27 may also inhibits the extrinsic pathway by blocking the Daxx2-mediated apoptosis [59].

Figure 3.31: Schematic representation of the inhibitory role of Hsp27 in cytochrome c-dependent activation of Cas-3 [31].

However, results achieved in this study, also showed that train (90 min) of short and long heat pulses are able to induce cellular damage as shown by the higher activation of the Casp-3 in spite of the higher phosphorylation of HSP27 compared to the CW heating. Small HSP exhibits temperature-dependent chaperone like-activity preventing the aggregation of stressed proteins [107]. In particular, it has been demonstrated by previous studies that HSP27 undergoes thermally induced self-association [108], leading to increased oligomeric size, which correlated with increase in its chaperone-like activity [109]. For instance, Garolla and Mauk [109] evaluates the chaperone activity of HSP27 as its ability to inhibit dithiothretoil-induced insulin aggregation as a function of the temperature, in the 20–48 °C temperature interval (Fig. 3.32).

2 After Fas stimulation, Daxx is activated playing its role of pro-apoptotic protein in activating the c-JUN- N-Terminal Kinase (JNK) pathway. 186

CHAPTER 3. Cell mortality and heat shock stress induced by pulsed modulated millimeter wave heating at cellular level

Results of this study showed interesting aspects of the HSP27 chaperone activity. HSP27 ability to protect cells increases with the temperature, exhibiting a sharp increase in the 34–43 °C temperature range. Specifically, the chaperone activity of HSP27 increases slightly between 20 °C and 30 °C (13.5% inhibition), with much greater increase observed between 34 °C and 43 °C (response increases by 1.4 times in this interval). At higher temperature, i.e., 43–48 °C, chaperone activity increases only slightly (4.6 %) exhibiting an apparent plateau. These results are consistent with the outcomes of the present study. The analysis of the dependence of the phosphorylation of HSP27 at pulsed MMW exposure on the distances from the center showed that in the region between 0–1.8 mm, the induction of HSP27 phosphorylation reaches maximal level exhibiting a plateau (Fig. 3.25b). This region corresponds to the peak temperature rise of 48–49 °C (Fig. 3.14b). These results suggest that the amplitude of thermal pulses plays an important role in inducing apoptosis in cells. Cells become more vulnerable to heat damage at peak temperatures within the 48–49°C interval. At these temperatures, in spite of the maximum chaperon activation, the cell protection mechanism based on activation of HSP27 phosphorylation and other heat shock proteins does not cope with increasing heat damage of cells.

Figure 3.32: Dependence of chaperone activity of HSP27 on temperature evaluated as inhibition of dithiothretoil-induced insulin aggregation [109].

Previous studies evidenced that, depending on the intensity of the stress, occurrence of phosphorylation induces the dissociation of large non-phosphorylated HSP27, consisting typically of about 24 subunits (i.e., 6 tetramers), in smaller rod-like oligomers which are probably tetrameric [64]. The rapid stress-induced phosphorylation is the result of stimulation of the p38 MAP kinase cascade and subsequent activation of MAPKAP kinases 2 and 3 which directly phosphorylate mammalian HSP27 at several distinct sites including Serine (Ser-15, -78, and -82 [110]) and Threonine (Thr-143) residues [42]. Interestingly, chaperone properties correlate with the ability of the protein to form large oligomers. As long as significant amounts of large HSP27 oligomers could

187

CHAPTER 3. Cell mortality and heat shock stress induced by pulsed modulated millimeter wave heating at cellular level be formed, in vitro chaperone properties (i.e., preventing thermal aggregation), are detected. Nevertheless, overexpression of phosphorylation of HSP27 down-regulates its chaperone activity by decreasing oligomerization of the protein and its consequent ability to trap and refold the thermally stressed proteins [64]. These observation seem to be consistent with the results of our study, demonstrating that greater phosphorylation of HSP27, probably associated with lower oligomerization of the proteins, results in the decrease of its chaperone-like activity as evidenced by the increase of the number of cells undergoing to apoptosis. An example of the cells located in central point of the exposure, aligned to the WG axis (i.e., picture number 1 of Fig. 3.8) is represented in Figure 3.33. The blue represents the cellular nucleus, while the red is the dye associated to the activation of Casp-3.

PW (1.5 s) PW (6 s)

CW Sham Fig.3.33: Fluorescence microscopy pictures of A375 cells for sham, CW, PW (1.5 s), and PW (6 s). Each 188

CHAPTER 3. Cell mortality and heat shock stress induced by pulsed modulated millimeter wave heating at cellular level picture is taken in the center of the exposure (i.e., picture number 1 of Fig. 3.8). Blue represents the nucleus, while the red colour represents the fluorescence due to cleavage of Casp-3.

3.6 Discussion

In this study, two aspects of the complex response of the A375 melanoma cells have been analysed, namely the heat shock response mediated by the phosphorylation of HSP27 and the associated cellular damage quantified as the number of cells undergoing to apoptosis. Cellular response induced by CW heating has been compared to PW heating with the same average temperature rise rate. Two different durations of the heat pulses have been considered, namely 1.5 s and 6 s. In order to quantify the dose delivered to the exposed cells, thermal dose has been calculated by using the methodologies described in the section 3.3.3 (Thermal dose). Results of our calculation are represented in the Table 3.2 as mean ± standard deviation over three temperature measurements. They are representative of the thermal dose in the center of the well, corresponding to the highest induced temperature reached in the cell monolayer.

Table 3.2. Thermal dose calculated for CW and PW (1.5 s and 6 s duration) as dm43 or CEM43°C Type of exposure dm43 [37] CEM 43°C [37] CW 0.68 ±0.05 23.9 ± 9.7 PW (1.5 s) 0.68 ±0.05 502.67 ± 128.42 PW (6 s) 0.80 ±0.1 1021.67 ± 401.5

Both models employed represent the integral over time of the temperature reached during the exposure in comparison to a certain threshold typically fixed at 43 °C [37]. The dm43 compares the average temperature to the initial temperature of the exposure (i.e., 32 °C), therefore no significant differences are observed among the conditions considered in the study. However, as expected, a little increase (15%) is observed for the long pulses compared to the other exposures carried out, due to the slightly higher average temperature rise observed in the temperature dynamics of long pulses (Fig. 3.19). The CEM43°C, which is extensively and successfully used in the practical clinic [111] allows to transform the time-temperature history of the treatment into an equivalent number of minutes of heating at 43°C. The model takes into account the different cellular sensitivity to heat as well as the different damage rate above and below the thermal threshold of 43 °C by means of the parameter R [38]. In terms of thermal dose, as expected, the dose level associated to exposure to long pulses is higher about the double of the thermal dose delivered by short pulses with the same 189

CHAPTER 3. Cell mortality and heat shock stress induced by pulsed modulated millimeter wave heating at cellular level average temperature dynamics. The main contribution for the achievement of such level of dose is given by temperatures in the interval between 47 °C and 49 °C. Particularly, more than a half of the dose is received by the cells when the peak temperature is above 48 °C. Interestingly, both phosphorylation of HSP27 and cellular apoptosis are higher when heat is given as short 1.5 s pulses. For instance, in the central area of the well (0 < d < 1.8 mm), cellular response of short pulses is 21% and 55% higher than long pulses in terms of thermal stress and cellular damage, respectively. Moreover, as the extension of the cellular response appear to be focused within a small area with radius of about 1.8 mm (Fig. 3.29), results suggest that the use of short pulses is advantageous because it may provide a high selectivity within the tissue accompanied to a dose level which is below the typical thermal threshold reported for human skin (i.e., around 600 CEM43°C [112]). Negligible cellular death following the CW exposure is consistent with the low value of thermal dose provided by this type of exposure. Results of this study suggest that minimal, mean, and peak temperatures are not the only parameters dominating the cellular response. For instance, temperature rise rate, duty cycle, energy as well as duration of the heat pulse may be among the parameters which mostly influence the cellular response. While not providing direct evidence in terms of cellular stress, several studies suggested that temperature rise rate could lead to changes at the cellular level. It was demonstrated that the transient heating (i.e., warm-up phase) up to several °C, at 75 GHz, leads to changes in the membrane potential and consequently in the firing rate of neurons [113], [114]. These changes were dependent on the temperature rise rate. Even small temperature rise rate of 0.0025 °C/s was found to be sufficient to record a transient inhibition of the firing rate of the snail neurons [113]. The phase of transient decrease of the firing rate could be singled out by varying the exposure pulse duration [114]. Another group demonstrated that temperature increase at 10 °C/s rate caused a temporary cessation in the firing of the pacemaker neurons [115]. Furthermore, it was reported that the viability of liver cancer cells exposed at 100 MHz depends on the temperature rise rate [116]. In this case, the cell viability was unaffected by temperature rise rates below 10 °C/s and decreased to 90% for the temperature rise rate of 50°C/s. Note that the effect was shown to be frequency dependent, and the cell mortality threshold shifted towards lower temperature rise rates when increasing the frequency to 2.45 GHz. Temperature rise rate of pulses used in this study is 6.7 °C/s and 1.7 °C/s for short and long pulses, respectively, and we cannot exclude that the cell damage also depends on it. Temperature rise rate of pulses with duration of 1.5 s is 75% higher than that of long pulses, resulting in a maximal cellular response, in the center of the well (0 < d < 0.5 mm) 28% and 57% higher than the one induced by long pulses, in terms of heat shock response and cellular 190

CHAPTER 3. Cell mortality and heat shock stress induced by pulsed modulated millimeter wave heating at cellular level damage, respectively. However, the effects of the amplitude of the pulse and temperature rise rate cannot be easily discriminated without further investigations. At present the mechanism of the direct action of temperature rise rate on the enzymatic reaction is not known and several hypothesis may explain the observed cellular responses. Thermal stimulation of a receptor induces activation of several downstream kinases, leading to the phosphorylation of the proteins. Phosphorylation/ dephosphorylation (see 3.4, Heat shock response) has been considered as a fundamental on/off switch for protein functions [117]. Faster temperature rise of the thermal train made by short pulses may be associated to faster phosphorylation/dephoshorylation of the proteins, possibly associated to a lower capability to trap and refold the proteins stressed by heat. On the contrary, train of long heat pulses, may be associated to slower phosphorylation/dephoshorylation process, allowing the cells to better exercise their function of molecular chaperones. Peak power of the exposure represents another parameter possibly determining the cellular response. In this study, faster short pulses of 1.5 s, associated to higher peak power (57%) compared to slower long pulses of 6 s, resulted in higher cellular response. Zorec et al. [118], in a study aimed to investigate transdermal drug delivery of compounds possessing different molecular weights by means of YAG laser, analysed the effect of duration, power, and energy of laser pulses. They found that the energy of the pulse mostly dictated the size/depth of laser-created microchannels, i.e., laser- ablation, while the duration of the pulses dictates the extent of thermally altered tissue. In particular, the study showed that shorter pulses with higher power allowed the achievement of higher ablative action, reducing unwanted thermal effects on tissue and increasing molecular delivery trough laser- created channels. When long pulses (250 μs) were used, the lowest deliver of the drug was observed, compared to shorter pulses (100 μs and 50 μs), even at the same delivered pulse energy. This is in agreements with results of our experiments where short pulses provided higher damage of cells. However, it is interesting to note that the energy associated to 1.5 s duration pulse is 46% lower than the energy of the long pulse (Table 3.1). The duty cycle, i.e., pulse duration over pulse period, also plays a role in defining the cellular response to heat pulses. In our study, heat pulses with a 65% lower duty cycle, as in the case of the short pulses, provided greater killing efficiency of the malignant cells, which is consistent with other previous studies [118], [119]. Pulse duration of the heat pulse seems to play a fundamental role in the determination of the spatial extension of the thermal stress as evidenced by the map of the fluorescence of cells reported in Figure 3.24. Long thermal pulses allowed heat to highly diffuse into the cells, providing a larger response in terms of phosphorylation of HSP27.

191

CHAPTER 3. Cell mortality and heat shock stress induced by pulsed modulated millimeter wave heating at cellular level

3.7 Conclusions In this study, for the first time, the cellular response of the A375 malignant melanoma cell line, following the exposure to CW and PW MMW-induced heating with the same average temperature dynamics is studied through high-content fluorescence microscopy image analysis. First, the use of MMW as a source of heating resulted in temperature rise of the cell monolayer confined into a small area within the center of the well (0 < d < 2.5 mm) where temperature variation is within the 1% and 4% for CW and PW exposures, respectively (Fig. 3.14). Results obtained in vitro support the use of MMW devices, operating within the 20–100 GHz band, for the spatially-accurate thermal treatment of in vivo melanoma, or in general superficial skin cancers. Second, the CW heating, within the moderate hyperthermic temperature range (e.g., around 42 °C), is not suitable for the thermal treatment of melanoma as it does not induce any noticeable apoptosis, which remains very close to the cellular baseline of the sham (i.e., about 1.5–2% of apoptotic cells are detected after CW heating). Third, short (e.g., 1.5 s) high-intensity heat pulses with the same average temperature dynamics as CW heating of about 42 °C and steady state peak temperature of about 49 °C, represent the optimum solution, among the exposures considered in the study, to obtain high-focused cellular apoptosis toward the location of the open-ended rectangular WG used as a source of heating. In the small area between 0 and 0.5 mm from the center of exposed well, short pulses induce an increase of the cellular apoptosis greater 2.3 and 5.2 times than that induced by long pulses and CW heating, respectively. Four, temperature dynamics, in terms of peak, average, and minimum temperature is not the only parameter playing a role in the determination of the cellular response as suggested by the fact that train of pulses (1.5 s and 6 s) with similar temperature dynamics resulted in a different cellular response. Temperature rise rate, peak power, duty cycle, energy as well as duration of the heat pulse may be among the parameters which determine the cellular response to heat pulses. Specifically, in this study, higher temperature rise rate of the short thermal pulses, associated to greater peak power and lower duty cycle and energy, resulted in stronger cellular responses compared to 6 s heat pulses. To conclude, this study represents a fundamental starting point towards the comprehension of the cellular response to MMW pulsed-induced heating, paving the way towards a new non-invasive strategy for local destruction of melanoma at low average heating. Optimization of the pulse shape, in terms of power, duty cycle, and temperature peak represents challenged for future studies in this direction with the scope to achieve the strongest localized cellular apoptosis in the minimally invasive way.

192

CHAPTER 3. Bibliography

Bibliography

[1] I. Mellal, A. Oukaira, E. Kengene, and A. Lakhssassi, “Thermal therapy modalities for cancer treatment: a review and future perspectives,” Int. J. Appl. Sci. - Res. Rev., vol. 4, no. 4, Nov. 2017. [2] D. A. Hodgson, I. B. Feldberg, N. Sharp, N. Cronin, M. Evans, and L. Hirschowitz, “Microwave endometrial ablation: development, clinical trials and outcomes at three years,” BJOG Int. J. Obstet. Gynaecol., vol. 106, no. 7, pp. 684–694, Jul. 1999. [3] H. Luyen, F. Gao, S. C. Hagness, and N. Behdad, “Microwave ablation at 10.0 GHz achieves comparable ablation zones to 1.9 GHz in ex vivo bovine liver,” IEEE Trans. Biomed. Eng., vol. 61, no. 6, pp. 1702–1710, Jun. 2014. [4] C. P. Hancock, N. Dharmasiri, M. White, and A. M. Goodman, “The design and development of an integrated multi-functional microwave antenna structure for biological applications,” IEEE Trans. Microw. Theory Tech., vol. 61, no. 5, pp. 2230–2241, May 2013. [5] J. Yoon et al., “High-frequency microwave ablation method for enhanced cancer treatment with minimized collateral damage,” Int. J. Cancer, vol. 129, no. 8, pp. 1970–1978, Oct. 2011. [6] M. Zhadobov, S. I. Alekseev, Y. Le Dréan, R. Sauleau, and E. E. Fesenko, “Millimeter waves as a source of selective heating of skin,” Bioelectromagnetics, vol. 36, no. 6, pp. 464–475, Sep. 2015. [7] B. S. Kalal, D. Upadhya, and V. R. Pai, “Chemotherapy resistance mechanisms in advanced skin cancer,” Oncol Rev, vol. 11, no. 1, p. 326, Mar. 2017. [8] E. K. Rofstad and T. Brustad, “Effect of hyperthermia on human melanoma cells heated either as solid tumors in athymic nude mice or in vitro,” Cancer, vol. 50, no. 7, pp. 1304–1308, 1982. [9] A. Ito et al., “Tumor regression by combined immunotherapy and hyperthermia using magnetic nanoparticles in an experimental subcutaneous murine melanoma,” Cancer Sci., vol. 94, no. 3, pp. 308–313, Mar. 2003. [10] M. Suzuki, M. Shinkai, H. Honda, and T. Kobayashi, “Anticancer effect and immune induction by hyperthermia of malignant melanoma using magnetite cationic liposomes,” Melanoma Res., vol. 13, no. 2, pp. 129–135, Apr. 2003. [11] E. M. Noorda, B. C. Vrouenraets, O. E. Nieweg, J. M. Klaase, J. vander Zee, and B. B. Kroon, “Long-term results of a double perfusion schedule using high dose hyperthermia and melphalan sequentially in extensive melanoma of the lower limb,” Melanoma Res., vol. 13, no. 4, p. 395, Aug. 2003. [12] J. Overgaard et al., “Hyperthermia as an adjuvant to radiation therapy of recurrent or metastatic malignant melanoma. A multicentre randomized trial by the European Society for Hyperthermic Oncology,” Int. J. Hyperth. Off. J. Eur. Soc. Hyperthermic Oncol. North Am. Hyperth. Group, vol. 12, no. 1, pp. 3–20, Feb. 1996. [13] T. Mantso et al., “Hyperthermia induces therapeutic effectiveness and potentiates adjuvant therapy with non-targeted and targeted drugs in an in vitro model of human malignant melanoma,” Sci. Rep., vol. 8, no. 1, p. 10724, Jul. 2018. [14] I. Szabo et al., “Destruction of cutaneous melanoma with millimeter wave hyperthermia in mice,” IEEE Trans. Plasma Sci., vol. 32, no. 4, pp. 1653–1660, Aug. 2004. [15] M. P. Garcia, J. R. T. Cavalheiro, and M. H. Fernandes, “Acute and long-term effects of hyperthermia in B16-F10 melanoma cells,” PLoS ONE, vol. 7, no. 4, p. e35489, Apr. 2012. [16] Y. G. Shellman et al., “Hyperthermia induces endoplasmic reticulum-mediated apoptosis in melanoma and non-melanoma skin cancer cells,” J. Invest. Dermatol., vol. 128, no. 4, pp. 949– 956, Apr. 2008. [17] A. B. Parrish, C. D. Freel, and S. Kornbluth, “Cellular mechanisms controlling caspase activation and function,” Cold Spring Harb. Perspect. Biol., vol. 5, no. 6, pp. a008672– a008672, Jun. 2013.

193

CHAPTER 3. Bibliography

[18] M. Bedoya, A. M. del Rio, J. Chiang, and C. L. Brace, “Microwave ablation energy delivery: Influence of power pulsing on ablation results in an ex vivo and in vivo liver model,” Med. Phys., vol. 41, no. 12, Dec. 2014. [19] J. W. Hunt, R. Lalonde, H. Ginsberg, S. Urchuk, and A. Worthington, “Rapid heating: critical theoretical assessment of thermal gradients found in hyperthermia treatments,” Int. J. Hyperth. Off. J. Eur. Soc. Hyperthermic Oncol. North Am. Hyperth. Group, vol. 7, no. 5, pp. 703–718, Oct. 1991. [20] C. S. Kim, R. P. Hill, J. C. Kumaradas, R. Irvine, F.-F. Liu, and M. D. Sherar, “Effect of simultaneous pulsed hyperthermia and pulsed radiation treatment on survival of SiHa cells,” Int. J. Hyperthermia, vol. 14, no. 6, pp. 573–581, Jan. 1998. [21] G. C. Li, N. F. Mivechi, and G. Weitzel, “Heat shock proteins, thermotolerance, and their relevance to clinical hyperthermia,” Int. J. Hyperthermia, vol. 11, no. 4, pp. 459–488, Jan. 1995. [22] R. I. Morimoto, “The heat shock response: systems biology of proteotoxic stress in aging and disease,” Cold Spring Harb. Symp. Quant. Biol., vol. 76, pp. 91–99, 2011. [23] D. M. Simanovskii et al., “Cellular tolerance to pulsed hyperthermia,” Phys. Rev. E Stat. Nonlin. Soft Matter Phys., vol. 74, no. 1 Pt 1, p. 011915, Jul. 2006. [24] C. D. Clark, W. J. Marshall, and R. J. Thomas, “Theoretical analysis of multiple-pulse thermal damage thresholds of the retina,” J. Laser Appl., vol. 25, no. 1, p. 012005, Jan. 2013. [25] M. Zhou, A. Zhang, B. Lin, J. Liu, and L. X. Xu, “Study of heat shock response of human umbilical vein endothelial cells (HUVECs) using cDNA microarray,” Int. J. Hyperth. Off. J. Eur. Soc. Hyperthermic Oncol. North Am. Hyperth. Group, vol. 23, no. 3, pp. 225–258, May 2007. [26] J. Nakayama, T. Kageshita, M. Nakashima, M. Tsujisaki, K. Imai, and Y. Hori, “Increase in shedding of intercellular adhesion molecule-1 in human malignant melanoma cell lines treated with hyperthermia in vitro,” Pigment Cell Res., vol. 9, no. 3, pp. 154–158, Jun. 1996. [27] V. Milani et al., “Melanoma-associated antigen tyrosinase but not Melan-A/MART-1 expression and presentation dissociate during the heat shock response,” Int. Immunol., vol. 17, no. 3, pp. 257–268, Jan. 2005. [28] R. S. Oliveira Filho, R. G. Bevilacqua, and R. Chammas, “Hyperthermia increases the metastatic potential of murine melanoma,” Braz. J. Med. Biol. Res. Rev. Bras. Pesqui. Medicas E Biol., vol. 30, no. 8, pp. 941–945, Aug. 1997. [29] ATCC, “A-375 [A375] ATCC ® CRL-1619TM Homo sapiens skin malignant melanoma.” [Online]. Available: http://www.lgcstandards-atcc.org/products/all/CRL- 1619.aspx?geo_country=fr. [30] C. Le Quément, C. N. Nicolaz, D. Habauzit, M. Zhadobov, R. Sauleau, and Y. Le Dréan, “Impact of 60-GHz millimeter waves and corresponding heat effect on endoplasmic reticulum stress sensor gene expression,” Bioelectromagnetics, vol. 35, no. 6, pp. 444–451, Sep. 2014. [31] P. Pandey et al., “Hsp27 functions as a negative regulator of cytochrome c-dependent activation of procaspase-3,” Oncogene, vol. 19, no. 16, pp. 1975–1981, Apr. 2000. [32] R. W. Burry, “Controls for immunocytochemistry: an update,” J. Histochem Cytochem, vol. 59, no. 1, pp. 6–12, Jan. 2011. [33] Novus Biologicals, “Immunocytochemistry (ICC) Handbook.” [Online]. Available: https://www.novusbio.com/support/immunocytochemistry-icc-handbook. [34] A. J. Haas, Y. Le Page, M. Zhadobov, R. Sauleau, and Y. Le Dréan, “Effects of 60-GHz millimeter waves on neurite outgrowth in PC12 cells using high-content screening,” Neurosci. Lett., vol. 618, pp. 58–65, Apr. 2016. [35] SelectScience, “Thermo Scientific ArrayScan® VTI HCS reader from Thermo Fisher Scientific.” [Online]. Available: https://www.selectscience.net/products/thermo-scientific- arrayscan-vti-hcs-reader/?prodID=80894.

194

CHAPTER 3. Bibliography

[36] M. Zhadobov, S. I. Alekseev, R. Sauleau, Y. Le Page, Y. Le Dréan, and E. E. Fesenko, “Microscale temperature and SAR measurements in cell monolayer models exposed to millimeter waves,” Bioelectromagnetics, vol. 38, no. 1, pp. 11–21, Jan. 2017. [37] A. S. Sapareto and W. C. Dewey, “Thermal dose determination in cancer therapy,” Int. J. Radiat. Oncol., vol. 10, no. 6, pp. 787–800, Apr. 1984. [38] J. A. Pearce, “Comparative analysis of mathematical models of cell death and thermal damage processes,” Int. J. Hyperthermia, vol. 29, no. 4, pp. 262–280, Jun. 2013. [39] A.-P. Arrigo, “Analysis of HspB1 (Hsp27) oligomerization and phosphorylation patterns and its interaction with specific client polypeptides,” Methods Mol. Biol. Clifton NJ, vol. 1709, pp. 163–178, 2018. [40] S. Dai et al., “Comprehensive characterization of heat shock protein 27 phosphorylation in human endothelial cells stimulated by the microbial dithiole thiolutin,” J. Proteome Res., vol. 7, no. 10, pp. 4384–4395, Oct. 2008. [41] F. Ardito, M. Giuliani, D. Perrone, G. Troiano, and L. L. Muzio, “The crucial role of protein phosphorylation in cell signaling and its use as targeted therapy (Review),” Int. J. Mol. Med., vol. 40, no. 2, pp. 271–280, Aug. 2017. [42] M. Katsogiannou, C. Andrieu, and P. Rocchi, “Heat shock protein 27 phosphorylation state is associated with cancer progression,” Front. Genet., vol. 5, Oct. 2014. [43] F. Plattner and J. A. Bibb, “Chapter 25 - Serine and Threonine phosphorylation,” in Basic Neurochemistry (Eighth Edition), S. T. Brady, G. J. Siegel, R. W. Albers, and D. L. Price, Eds. New York: Academic Press, 2012, pp. 467–492. [44] T. A. Hamilton, “Protein kinases,” in Encyclopedia of Immunology (Second Edition), P. J. Delves, Ed. Oxford: Elsevier, 1998, pp. 2028–2033. [45] L. N. Johnson and R. J. Lewis, “Structural basis for control by phosphorylation,” Chem. Rev., vol. 101, no. 8, pp. 2209–2242, Aug. 2001. [46] P. Cohen, “The regulation of protein function by multisite phosphorylation--a 25 year update,” Trends Biochem. Sci., vol. 25, no. 12, pp. 596–601, Dec. 2000. [47] N. K. Tonks, “Protein tyrosine phosphatases: from genes, to function, to disease,” Nat. Rev. Mol. Cell Biol., vol. 7, no. 11, pp. 833–846, Nov. 2006. [48] S. Kostenko and U. Moens, “Heat shock protein 27 phosphorylation: kinases, phosphatases, functions and pathology,” Cell. Mol. Life Sci., vol. 66, no. 20, pp. 3289–3307, Oct. 2009. [49] M. J. Vos, B. Kanon, and H. H. Kampinga, “HSPB7 is a SC35 speckle resident small heat shock protein,” Biochim. Biophys. Acta, vol. 1793, no. 8, pp. 1343–1353, Aug. 2009. [50] J. Landry, P. Chrétien, H. Lambert, E. Hickey, and L. A. Weber, “Heat shock resistance conferred by expression of the human HSP27 gene in rodent cells.,” J. Cell Biol., vol. 109, no. 1, pp. 7–15, Jan. 1989. [51] A. L. Yu, R. Fuchshofer, M. Birke, A. Kampik, H. Bloemendal, and U. Welge-Lüssen, “Oxidative stress and TGF-beta2 increase heat shock protein 27 expression in human optic nerve head astrocytes,” Invest. Ophthalmol. Vis. Sci., vol. 49, no. 12, pp. 5403–5411, Dec. 2008. [52] C. K. Lee, R. Weindruch, and T. A. Prolla, “Gene-expression profile of the ageing brain in mice,” Nat Genet, vol. 25, no. 3, pp. 294–297, Jul. 2000. [53] D. R. Ciocca and S. K. Calderwood, “Heat shock proteins in cancer: diagnostic, prognostic, predictive, and treatment implications,” Cell Stress Chaperones, vol. 10, no. 2, pp. 86–103, Jun. 2005. [54] T. F. Outeiro et al., “Small heat shock proteins protect against alpha-synuclein-induced toxicity and aggregation,” Biochem. Biophys. Res. Commun., vol. 351, no. 3, pp. 631–638, 2006. [55] R. A. Stetler et al., “Phosphorylation of HSP27 by protein kinase D is essential for mediating neuroprotection against ischemic neuronal injury,” J. Neurosci. Off. J. Soc. Neurosci., vol. 32, no. 8, pp. 2667–2682, Feb. 2012. 195

CHAPTER 3. Bibliography

[56] J. Guay, H. Lambert, G. Gingras-Breton, J. N. Lavoie, J. Huot, and J. Landry, “Regulation of actin filament dynamics by p38 map kinase-mediated phosphorylation of heat shock protein 27,” J. Cell Sci., vol. 110 ( Pt 3), pp. 357–368, Feb. 1997. [57] A.-P. Arrigo and B. Gibert, “HspB1 dynamic phospho-oligomeric structure dependent interactome as cancer therapeutic target,” Curr. Mol. Med., vol. 12, no. 9, pp. 1151–1163, Nov. 2012. [58] T. Watanabe, S. Sekine, I. Naguro, Y. Sekine, and H. Ichijo, “Apoptosis signal-regulating Kinase 1 (ASK1)-p38 pathway-dependent cytoplasmic translocation of the orphan nuclear receptor NR4A2 is required for oxidative stress-induced necrosis,” J. Biol. Chem., vol. 290, no. 17, pp. 10791–10803, Apr. 2015. [59] S. J. Charette, J. N. Lavoie, H. Lambert, and J. Landry, “Inhibition of Daxx-mediated apoptosis by heat shock protein 27,” Mol. Cell. Biol., vol. 20, no. 20, pp. 7602–7612, Oct. 2000. [60] J. M. Bruey et al., “Hsp27 negatively regulates cell death by interacting with cytochrome c,” Nat. Cell Biol., vol. 2, no. 9, pp. 645–652, Sep. 2000. [61] A. Havasi et al., “Hsp27 inhibits bax activation and apoptosis via a phosphatidylinositol 3- kinase-dependent mechanism,” J. Biol. Chem., vol. 283, no. 18, pp. 12305–12313, May 2008. [62] A. Samali et al., “Hsp27 protects mitochondria of thermotolerant cells against apoptotic stimuli,” Cell Stress Chaperones, vol. 6, no. 1, pp. 49–58, Jan. 2001. [63] C. Paul, S. Simon, B. Gibert, S. Virot, F. Manero, and A.-P. Arrigo, “Dynamic processes that reflect anti-apoptotic strategies set up by HspB1 (Hsp27),” Exp. Cell Res., vol. 316, no. 9, pp. 1535–1552, May 2010. [64] T. Rogalla et al., “Regulation of Hsp27 oligomerization, chaperone function, and protective activity against oxidative stress/tumor necrosis factor α by phosphorylation,” J. Biol. Chem., vol. 274, no. 27, pp. 18947–18956, Feb. 1999. [65] P. Mehlen, E. Hickey, L. A. Weber, and A.-P. Arrigo, “Large unphosphorylated aggregates as the active form of hsp27 which controls intracellular reactive oxygen species and glutathione levels and generates a protection against TNFα in NIH-3T3-ras cells,” Biochem. Biophys. Res. Commun., vol. 241, no. 1, pp. 187–192, Dec. 1997. [66] E. K. Rofstad and T. Brustad, “Effect of hyperthermia on human melanoma cells heated either as solid tumors in athymic nude mice or in vitro,” Cancer, vol. 50, no. 7, pp. 1304–1308, Oct. 1982. [67] H. I. Robins et al., “Whole body hyperthermia (41–42 degrees C): a simple technique for unanesthetized mice,” Med. Phys., vol. 11, no. 6, pp. 833–839, Dec. 1984. [68] A. M. Westermann et al., “First results of triple-modality treatment combining radiotherapy, chemotherapy, and hyperthermia for the treatment of patients with stage IIB, III, and IVA cervical carcinoma,” Cancer, vol. 104, no. 4, pp. 763–770, Aug. 2005. [69] S. Jacques, “How tissue optics affect dosimetry for photochemical, photothermal, and photomechanical mechanisms of laser-tissue interaction,” in Proceedings of SPIE - The International Society for Optical Engineering, 1992, pp. 316–322. [70] Thermo Fisher Scientific, “SYTOX® Green nucleic acid stain protocol.” [Online]. Available: https://www.thermofisher.com/fr/fr/home/references/protocols/cell-and-tissue- analysis/protocols/sytox-green-protocol.html. [71] Thermo Fisher Scientific, “LIVE/DEAD Cell viability assays.” [Online]. Available: https://www.thermofisher.com/fr/fr/home/brands/molecular-probes/key-molecular-probes- products/live-dead-viability-brand-page.html. [72] S. Uesugi, K. Yamashita, K. Nakashima, and H. Ito, “Apoptotic cell death induced by local brain hyperthermia in a rat glioma model,” Acta Neuropathol. (Berl.), vol. 96, no. 4, pp. 351– 356, Sep. 1998. [73] K. Ahmed, Y. Tabuchi, and T. Kondo, “Hyperthermia: an effective strategy to induce apoptosis in cancer cells,” Apoptosis, vol. 20, no. 11, pp. 1411–1419, Nov. 2015.

196

CHAPTER 3. Bibliography

[74] A. Peña‐Blanco and A. J. García‐Sáez, “Bax, Bak and beyond — mitochondrial performance in apoptosis,” FEBS J., vol. 285, no. 3, pp. 416–431, 2018. [75] D. W. Nicholson et al., “Identification and inhibition of the ICE/CED-3 protease necessary for mammalian apoptosis,” Nature, vol. 376, no. 6535, pp. 37–43, Jul. 1995. [76] Novus Biologicals, “Caspase-3 activation - An indicator of apoptosis in image-based assays.” [Online]. Available: http://www.bio-medicine.org/biology-technology/Caspase-3- Activation---An-Indicator-of-Apoptosis-in-Image-Based-Assays-1346-1/. [77] S. Fulda and K.-M. Debatin, “Extrinsic versus intrinsic apoptosis pathways in anticancer chemotherapy,” Oncogene, vol. 25, no. 34, pp. 4798–4811, Aug. 2006. [78] M. O. Hengartner, “The biochemistry of apoptosis,” Nature, vol. 407, no. 6805, pp. 770– 776, Oct. 2000. [79] Y. Shi, “Caspase activation, inhibition, and reactivation: A mechanistic view,” Protein Sci., vol. 13, no. 8, pp. 1979–1987, Aug. 2004. [80] E. S. Alnemri et al., “Human ICE/CED-3 Protease Nomenclature,” Cell, vol. 87, no. 2, p. 171, Oct. 1996. [81] K. Matsuura, K. Canfield, W. Feng, and M. Kurokawa, “Metabolic regulation of apoptosis in cancer,” in International Review of Cell and Molecular Biology, vol. 327, Elsevier, 2016, pp. 43–87. [82] R. C. Taylor, S. P. Cullen, and S. J. Martin, “Apoptosis: controlled demolition at the cellular level,” Nat. Rev. Mol. Cell Biol., vol. 9, no. 3, pp. 231–241, Mar. 2008. [83] S. Nagata, “Apoptotic DNA fragmentation,” Exp. Cell Res., vol. 256, no. 1, pp. 12–18, Apr. 2000. [84] S. Elmore, “Apoptosis: a review of programmed cell death,” Toxicol. Pathol., vol. 35, no. 4, pp. 495–516, Jun. 2007. [85] R. M. Locksley, N. Killeen, and M. J. Lenardo, “The TNF and TNF receptor superfamilies: integrating mammalian biology,” Cell, vol. 104, no. 4, pp. 487–501, Feb. 2001. [86] G. M. Cooper, “Cell-Cell Interactions,” Cell Mol. Approach 2nd Ed., 2000. [87] H. Walczak and P. H. Krammer, “The CD95 (APO-1/Fas) and the TRAIL (APO-2L) Apoptosis Systems,” Exp. Cell Res., vol. 256, no. 1, pp. 58–66, Apr. 2000. [88] F. C. Kischkel et al., “Cytotoxicity-dependent APO-1 (Fas/CD95)-associated proteins form a death-inducing signaling complex (DISC) with the receptor,” EMBO J., vol. 14, no. 22, pp. 5579–5588, Nov. 1995. [89] K. M. Boatright et al., “A unified model for apical Caspase activation,” Mol. Cell, vol. 11, no. 2, pp. 529–541, 2003. [90] I. Budihardjo, H. Oliver, M. Lutter, X. Luo, and X. Wang, “Biochemical pathways of caspase activation during apoptosis,” Annu. Rev. Cell Dev. Biol., vol. 15, no. 1, pp. 269–290, Nov. 1999. [91] H. Li, H. Zhu, C.-J. Xu, and J. Yuan, “Cleavage of BID by Caspase 8 mediates the mitochondrial damage in the Fas pathway of apoptosis,” Cell, vol. 4, no. 4, pp. 491–521, Aug. 1998. [92] D. Westphal, G. Dewson, P. E. Czabotar, and R. M. Kluck, “Molecular biology of Bax and Bak activation and action,” Biochim. Biophys. Acta BBA - Mol. Cell Res., vol. 1813, no. 4, pp. 521–531, Apr. 2011. [93] R & D Systems, “Apoptosis II: beyond cytochrome c; the other mitochondrial proteins:,” 2004. [Online]. Available: https://www.rndsystems.com/resources/articles/apoptosis-ii-beyond- cytochrome-c-other-mitochondrial-proteins. [94] X. Saelens, N. Festjens, L. V. Walle, M. van Gurp, G. C. van Loo, and P. Vandenabeele, “Toxic proteins released from mitochondria in cell death,” Oncogene, vol. 23, pp. 2861–2874, 2004.

197

CHAPTER 3. Bibliography

[95] K. Cain et al., “Apaf-1 oligomerizes into biologically active approximately 700-kDa and inactive approximately 1.4-MDa apoptosome complexes,” J. Biol. Chem., vol. 275, no. 9, pp. 6067–6070, Mar. 2000. [96] E. D. Crawford and J. A. Wells, “Caspase substrates and cellular remodeling,” Annu. Rev. Biochem., vol. 80, pp. 1055–1087, Jul. 2011. [97] M. N. Rylander, Y. Feng, J. Bass, and K. R. Diller, “Thermally induced injury and heat- shock protein expression in cells and tissues,” Ann. N. Y. Acad. Sci., vol. 1066, no. 1, pp. 222– 242, Mar. 2006. [98] A. Nussenzweig, P. Burgman, and G. C. Li, “The role of heat shock proteins in thermotolerance,” in Advances in Molecular and Cell Biology, vol. 19, Elsevier, 1997, pp. 261– 285. [99] S. A. Sapareto, L. E. Hopwood, W. C. Dewey, M. R. Raju, and J. W. Gray, “Effects of hyperthermia on survival and progression of Chinese Hamster Ovary cells,” Cancer Res., vol. 38, no. 2, pp. 393–400, Feb. 1978. [100] J. R. Lepock and J. Kruuv, “Thermotolerance as a possible cause of the critical temperature at 43 degrees in mammalian cells,” Cancer Res., vol. 40, no. 12, pp. 4485–4488, Dec. 1980. [101] H. M. Beere et al., “Heat-shock protein 70 inhibits apoptosis by preventing recruitment of procaspase-9 to the Apaf-1 apoptosome,” Nat. Cell Biol., vol. 2, no. 8, pp. 469–475, Aug. 2000. [102] M. A. Mackanos and C. H. Contag, “Pulse duration determines levels of Hsp70 induction in tissues following laser irradiation,” J Biomed Opt, vol. 16, no. 7, p. 078002, Jul. 2011. [103] G. M. Hahn and G. C. Li, “Thermotolerance and heat shock proteins in mammalian cells,” Radiat. Res., vol. 92, no. 3, pp. 452–457, Dec. 1982. [104] C. Garrido, J.-M. Bruey, A. Fromentin, A. Hammann, A. P. Arrigo, and E. Solary, “HSP27 inhibits cytochrome c -dependent activation of procaspase-9,” FASEB J., vol. 13, no. 14, pp. 2061–2070, Nov. 1999. [105] H. Konishi et al., “Activation of protein kinase B (Akt/RAC-protein kinase) by cellular stress and its association with heat shock protein Hsp27,” FEBS Lett., vol. 410, no. 2–3, pp. 493–498, 1997. [106] S. G. Kennedy, E. S. Kandel, T. K. Cross, and N. Hay, “Akt/Protein kinase B inhibits cell death by preventing the release of cytochrome c from mitochondria,” Mol. Cell. Biol., vol. 19, no. 8, pp. 5800–5810, Aug. 1999. [107] R. Bakthisaran, R. Tangirala, and C. M. Rao, “Small heat shock proteins: Role in cellular functions and pathology,” Biochim. Biophys. Acta BBA - Proteins Proteomics, vol. 1854, no. 4, pp. 291–319, Apr. 2015. [108] B. Lelj-Garolla and A. G. Mauk, “Self-association of a small heat shock protein,” J. Mol. Biol., vol. 345, no. 3, pp. 631–642, Jan. 2005. [109] B. Lelj-Garolla and A. G. Mauk, “Self-association and chaperone activity of Hsp27 are thermally activated,” J. Biol. Chem., vol. 281, no. 12, pp. 8169–8174, Mar. 2006. [110] D. Hayes, V. Napoli, A. Mazurkie, W. F. Stafford, and P. Graceffa, “Phosphorylation dependence of hsp27 multimeric size and molecular chaperone function,” J. Biol. Chem., vol. 284, no. 28, pp. 18801–18807, Jul. 2009. [111] G. C. van Rhoon, “Is CEM43 still a relevant thermal dose parameter for hyperthermia treatment monitoring?,” Int. J. Hyperth. Off. J. Eur. Soc. Hyperthermic Oncol. North Am. Hyperth. Group, vol. 32, no. 1, pp. 50–62, 2016. [112] G. C. van Rhoon, T. Samaras, P. S. Yarmolenko, M. W. Dewhirst, E. Neufeld, and N. Kuster, “CEM43°C thermal dose thresholds: a potential guide for magnetic resonance radiofrequency exposure levels?,” Eur. Radiol., vol. 23, no. 8, pp. 2215–2227, Aug. 2013. [113] S. I. Alekseev, M. C. Ziskin, N. V. Kochetkova, and M. A. Bolshakov, “Millimeter waves thermally alter the firing rate of the Lymnaea pacemaker neuron,” Bioelectromagnetics, vol. 18, no. 2, pp. 89–98, 1997.

198

CHAPTER 3. Bibliography

[114] S. I. Alekseev, M. S. Ziskin, and N. V. Kochetkova, “Effects of millimeter wavelength electromagnetic radiation on neurons: electrophysiological study,” Crit. Rev. Biomed. Eng., vol. 28, no. 5–6, pp. 52–59, 2000. [115] K. R. Foster, “Thermal and nonthermal mechanisms of interaction of radio-frequency energy with biological systems,” IEEE Trans. Plasma Sci., vol. 28, no. 1, pp. 15–23, Feb. 2000. [116] K. H. Schoenbach and S. Xiao, “Method and system for treating a biological target region using pulsed electromagnetic radiation,” US20120089209A1, 12-Apr-2012. [117] C. Salazar and T. Höfer, “Multisite protein phosphorylation – from molecular mechanisms to kinetic models,” FEBS J., vol. 276, no. 12, pp. 3177–3198, Jun. 2009. [118] B. Zorec, D. Škrabelj, M. Marinček, D. Miklavčič, and N. Pavšelj, “The effect of pulse duration, power and energy of fractional Er:YAG laser for transdermal delivery of differently sized FITC dextrans,” Int. J. Pharm., vol. 516, no. 1–2, pp. 204–213, Jan. 2017. [119] J. Tu et al., “Controllable in vivo hyperthermia effect induced by pulsed high intensity focused ultrasound with low duty cycles,” Appl. Phys. Lett., vol. 101, no. 12, p. 124102, Sep. 2012. [120] M. Bedoya, A. M. del Rio, J. Chiang, and C. L. Brace, “Microwave ablation energy delivery: Influence of power pulsing on ablation results in an ex vivo and in vivo liver model,” Med. Phys., vol. 41, no. 12, Dec. 2014.

199

General conclusions

In this work, for the first time the possibility to use and control MMW-induced heating as a means to thermally alter cancer cells has been studied. The main objective of the work was to provide an insight into understanding of modifications induced in the A375 malignant melanoma cell line by continuous (CW) and pulsed modulated (PW) MMW-induced heating with the same average temperature rise, at 58.4 GHz. First, thermodynamic aspects, often disregarded during in vitro bioelectromagnetic experiments, were addressed with the aim to extend the current knowledge about thermal dosimetry in in vitro experiments at MMW, towards the comprehension of the possible effects of the convective currents on cells. Specifically, we analysed in detail convection in a typical in vitro scenario at MMW. Experiments were carried out by using a specific exposure system made up by i) an exposure chamber, ii) a signal generator sub-unit, and iii) a temperature interface. Temperature measurements of the electromagnetically induced heating were performed in the samples following the methodology presented and validated by a previous study carried out by the WAVES team of the IETR by means of micro-thermocouples [1]. Local temperature dynamics were investigated in different liquids (i.e., distilled water, culture medium, and a cell monolayer covered by culture medium) and compared to the ones recorded in a convection-free water-equivalent model, i.e., agar gel. The results suggested that in planning in vitro studies, convection, triggered by a highly non- uniform distribution of SAR, caused by the shallow penetration of MMW, is one of the parameters that has to be carefully taken into account. It may indirectly impact the cellular response through effects, which are absent in the sham control, including local change of the concentration of oxygen or nutrients transported by the culture medium. The study revealed several important features of the development of convection in liquids exposed to 58.4 GHz: 1) a thin cell monolayer (i.e., 5–10 μm thick) does not impact the temperature dynamics in the culture medium, hence it does not influence the initiation of convection, 2) the onset of convection is preceded by the appearance of a temperature peak, which occurrence in time depends on SAR (the higher the SAR, the earlier the convection is triggered), 3) occurrence of convection depends on the liquid volume (larger volumes result in earlier triggering of convection and greater cooling effect), 4) initiation of convection depends on the viscosity of the medium (higher values of viscosity result in weaker convection), and 5) convection strongly changes the heat pulse shape decreasing the peak temperature and increasing the cooling rate. Second, the cellular response of the A375 melanoma cell line was characterized following the exposure to CW and PW MMW-induced heating, with the same average temperature dynamics.

200

General conclusions

Different durations of pulses, specifically 1.5 s and 6 s, with the same minimum, peak, and average temperature dynamics were compared. Two fundamental aspects defining the thermal response of the cells were analysed, including the phosphorylation of a heat shock protein (e.g., HSP27) as an indicator the cellular stress, and the activation of a specific protease enzyme (Caspase-3) playing a fundamental role in apoptosis. High-content fluorescence microscopy image analysis allowed us to follow the spatial variation of the biological response. Results evidenced that the CW heating provided at temperature lower than 43°C is not sufficient to induce cellular apoptosis in the malignant melanoma cells, thus being not suitable for the thermo-oncological treatment of superficial skin cancers. The distribution of the phosphorylation of HSP27 suggests that a thermotolerance response was induced in the exposed cells resulting in the effective inhibition of the activation of Caspase-3. On the contrary, high intensity heat pulses with peak amplitude of the single pulse of 10°C in the center of the well, and steady-state temperature peak of about 49°C, were able to induce a stronger cellular response compared to the one induced by CW heating, both in terms of phosphorylation of HSP27 and activation of Caspase-3. In particular, in the area up to 1.8 mm from the center of the well (with SAR, ΔT [PW max], and ΔT [CW] values within the 77%, 86%, and 96% in respect to the peak in the center, respectively), cellular apoptosis observed after the exposure to thermal pulses showed a drastic increase, consistent with the fact that the refolding system is saturated as suggested by the presence of a plateau in the phosphorylation of HSP27 in the same region, i.e., phosphorylated HSP27 is no more able to trap and refold thermal stressed proteins leading to cell death. Short heat pulses of 1.5 s provided greater cellular response. Specifically, in the area between 0 and 0.5 mm from the WG axis, phosphorylation of HSP27 induced by thermal pulses of 1.5 s was 1.4 and 1.2 times higher than the one induced by heat pulses of 6 s and CW- induced heating, respectively. The corresponding activation of the cleaved Caspase-3 induced by 1.5 s heat pulses was 2.3 and 5.2 times greater than that induced by 6 s pulses and CW heating, respectively. Duration of the heat pulse played a major role in the definition of the extension of the area of the thermal stress, i.e., the higher the duration of the thermal pulse, the larger the extent of the response of cells as visualized by the map of the induction of the phosphorylation of HSP27.

201

Future work and pespectives

Future work and perspectives

The analysis of the occurrence of convection in liquids exposed to CW and PW MMW heating was carried out through an experimental procedure involving the use of micro-thermocouple with leads diameter of 75 μm [1] allowing local measurement of the temperature rise. However, the complete understanding of thermodynamic events occurring in in vitro experiments at MMW, would highly benefit from measurements of the temperature distribution within the whole area of interest. Methodologies allowing the 2–D or 3–D mapping of the temperature would provide great improvement in the understanding of the development of convective currents during in vitro bioelectromagnetics experiments. The possibility to use thermochromic liquid crystals, TLC, (1.2.5, Techniques for measurement of MMW-induced heating ) appears as an attractive solution suitable for the 2–D visualization of the convective flows [2] as well as for the mapping of the temperature rise in the area of interest. To this end, a preliminary study to evaluate the feasibility of the experiments has been carried out. Among the different TLC mixtures currently commercially available [3] both microencapsulated crystals (slurries and coatings) and coated printed sheets appear the most convenient solutions. Moreover, 3–D temperature visualization by means of magnetic resonance thermal imaging (MRTI) is also envisaged to non-invasively map the temperature distribution of the exposed object as suggested by previous studies performed by the group of Alon et al. [4], [5]. The high spatial resolution of the MRTI technique would allow to clarify some of the empirical phenomena observed in our experiments, such as the formation of a temperature peak preceding the initiation of convection or the highest temperature rise observed in water, compared to agar gel, during PW exposure, for durations of the heat pulse between 2.5 and 4 s. Finally, as the work was mainly focused on the experimental study of the thermal convection during in vitro exposures, numerical analysis of multi-physics problem combining the electromagnetic and thermodynamic aspects will be considered for deeper and complete investigation of convection mechanisms. The response of the A375 human malignant melanoma cell line was investigated for the first time in this study following the exposure to MMW-induced heat pulses of different durations. The elicited effects were explored by means of immunofluorescence protocols targeting phosphorylated HSP27 and cleaved Caspase-3. Further investigations are needed for complete comprehension of the bio-physical phenomena behind the induced cellular death. The exact role of the parameters characterizing the pulse including peak power, duty cycle, or temperature rise rate is still unclear. Future investigations are oriented towards the application of pulses of different durations and shapes to discriminate the role of each parameter and possibly get insight into the exact mechanism

202

Future work and pespectives of heat-induced damage. Moreover, the understanding of the pathway triggered by thermal pulses to lead cells to apoptosis represents another future direction of the study. The real-time measurement of the temperature at the single-cell level is also envisaged. The temperature-dependent fluorescence property of Rhodamine-B appears extremely attractive for the evaluation of the real time temperature changes in the exposed cells as suggested by some previous studies [6], [7]. The calculation of the kinetics parameters, namely the frequency factor A ( ) and the activation energy ( ) (for more detail refer to the sub-section Kinetics of thermal damage process, in section1.3.2, Thermal dose) represents another perspective of the study. This would allow to take into account of the specific cellular sensitivity of the cell line considered, leading to the correct characterization of the parameter R used to calculate the CEM43 °C [8]. To this end, survival curves of the malignant melanoma cells will be provided, by using the protocol described in the section 1.3.2 (sub- section CEM43°C). Preliminary studies have been performed in order to evaluate the cellular sensitivity to continuous exposure to 44 °C, 46 °C, and 48 °C at different exposure durations ranging from 1 to 3 h. However, further validations of results are needed. Moreover, cellular sensitivity to heat pulses of different durations and exposure time will be also analysed. Finally, the development of a new mathematical model of thermal dose for pulsed heating is envisaged with the aim to improve the cellular damage prediction in the case of pulsed thermal therapies.

203

Bibliography

[1] M. Zhadobov, S. I. Alekseev, R. Sauleau, Y. Le Page, Y. Le Dréan, and E. E. Fesenko, “Microscale temperature and SAR measurements in cell monolayer models exposed to millimeter waves,” Bioelectromagnetics, vol. 38, no. 1, pp. 11–21, Jan. 2017. [2] T. Kowalewski, “Particle image velocimetry and thermometry using liquid crystal tracers,” presented at the 4th International Symposium on Particle Image Velocimetry, Göttingen, Germany, 2001. [3] Hallcrest, LCR, Glenview, “Handbook of thermochromic liquid crystal technology.” 2014. [4] L. Alon, S. Gabriel, G. Y. Cho, R. Brown, and C. M. Deniz, “Prospects for millimeter-wave compliance measurement technologies [Measurements Corner],” IEEE Antennas Propag. Mag., vol. 59, no. 2, pp. 115–125, Apr. 2017. [5] L. Alon, D. K. Sodickson, and C. M. Deniz, “Heat equation inversion framework for average SAR calculation from magnetic resonance thermal imaging,” Bioelectromagnetics, vol. 37, no. 7, pp. 493–503, Oct. 2016. [6] D. Moreau, C. Lefort, R. Burke, P. Leveque, and R. P. O’Connor, “Rhodamine B as an optical thermometer in cells focally exposed to infrared laser light or nanosecond pulsed electric fields,” Biomed Opt Express, vol. 6, no. 10, pp. 4105–4117, Oct. 2015. [7] S. Kohler, R. P. O’Connor, T. D. T. Vu, P. Leveque, and D. Arnaud-Cormos, “Experimental microdosimetry techniques for biological cells exposed to nanosecond pulsed electric fields using microfluorimetry,” IEEE Trans. Microw. Theory Tech., vol. 61, no. 5, pp. 2015–2022, May 2013. [8] S. A. Sapareto and W. C. Dewey, “Thermal dose determination in cancer therapy,” Int. J. Radiat. Oncol. Biol. Phys., vol. 10, no. 6, pp. 787–800, 1984.

204

Appendices

Content

Appendix A Millimeter-wave interaction with eye ...... 207

Appendix B Measurement of complex permittivity ...... 211 B.1 Calibration procedure ...... 212 B.2 Results ...... 213

Appendix C Measurement of thermal properties ...... 217 C.1 Measurement of thermal conductivity ...... 217 C.2 Measurement of heat capacity ...... 220

Appendix D Measurement of viscosity ...... 225

Appendix A. Millimeter-wave interaction with eye

Appendix A

Millimeter-wave interaction with eye The aim of this Appendix is to provide a non-exhaustive review of the main work done thus far about the interaction between MMW and eyes. Along with skin, eyes represent a major target for the MMW. Exposure of the eyes leads to the absorption of the MMW energy by the cornea (Fig. A.1) characterized by a free water content of 75% and a thickness of 0.5 mm [1]. Unlike other parts of the body, such as hands and arms, the blood flow in the eyes is not sufficient to allow the easy dissipation of the heat [2]. Thus, care should be taken to prevent unsafe overexposure of the eyes. Several studies have been performed in the MMW frequency band such as the ones at 1) 35 GHz [3], [4], 2) 40 GHz [5], 3) 60 GHz [6]–[8], 4) 75 GHz [5], 5) 76 GHz [9], 6) 94 GHz [3], 7) 95 GHz [5], and 8) 107 GHz [4]. A study performed by Kues et al. [6] did not report any detectable physiological ocular damage on the rabbit eyes exposed at 60 GHz at 10 mW/ cm2 for 8h. However, it has been demonstrated that MMW exposures at higher intensities produce adverse effects on eyes. For instance, authors in [7] identified thermal ocular lesions (e.g., corneal edema and desiccation) at 60 GHz after high intensity exposure (i.e., 1898 mW/ cm2, 6 min). In vivo exposures to continuous 200, 100, 75, 50, 10 mW/ cm2 at 76 GHz for 6 min, with a lens antenna revealed [9]: 1) corneal opacity, epithelial injury, miosis, and ocular inflammation up to 2–3 days after exposure at 200 mW/ cm2, 2) no ocular changes other than reversible corneal epithelial injury were seen following exposures at 100 and 75 mW/ cm2, and 3) no ocular changes after exposure at doses of 50 or 10 mW/ cm2. More recently, in vivo studies carried out at 40, 75, and 95 GHz [5] at two different durations (i.e., 6 and 30 min) assessed that 1) ocular damage was independent from the frequency of the exposure, 2) the temperature increase was frequency dependent (i.e., increased in the order 40 < 95 < 75 GHz), 3) the dose level to cause ocular damage with 50% probability, was power dependent and in the order 40 GHz (206 mW/ ) > 95 GHz (146 mW/ ) ≈ 75 GHz (143 mW/ ), and 4) for the same power density, ocular damage was greater after long exposures. Chalfin et al. [3] studied the effects of pulsed (1.5–5 s) 35 and 94 GHz exposure on the anterior segment of the non-human primate eye at the exposure intensities of 2 and 8 W/cm2. The authors

207

Appendix A. Millimeter-wave interaction with eye

detected distinct lesions, involving only superficial layers of the cornea. The corneal damage caused by 35 GHz exposure at 2 W/cm2 for 1.5–5 s was reversible within 24 h. The mean thresholds to induce damage at 35 GHz and 94 GHz were respectively 7.5 J/cm2 and 5 J/cm2 Rosenthal et al. [4], reported that continuous 35 or 107 GHz exposures at 50 mW/ cm2 for 15–80 min induced both epithelial damage and stromal edema in the cornea of rabbit eyes, although they began to recover the next day. Moreover, several numerical models have been developed to describe the temperature distribution in rabbit [8], non-human primate [10], and human eyes [11], [12] following MMW exposure. The model developed by Papaioannu and Samaras [8] takes into account the fluid dynamics of the aqueous humor (i.e., the blood surrogate for the eye, providing nutrition as well as contributing to the regulation of the homeostasis of ocular tissues) and the realistic boundary conditions implemented at the interface between the cornea and the environment. The model includes only the frontal part of the eye, designed using 3–D schematic components (i.e., the exact anatomical structure is not represented). Results of the study showed that at 60 GHz, for an incident power density of 475 mW/cm2, the maximum temperature predicted was of 45.8 °C, which is in good agreement with experimental results reported in [7]. In addition, the temperature gradient between the cornea and the iris determines a flow of the aqueous humor in the anterior chamber, which direction can be inverted for incident power density above 100 mW/cm2. Foster et al., [10] applied a 1–D heat conduction equation to the experimental data obtained by Chalfin et al. [3]. The thresholds for damage to the cornea corresponded to temperature elevation by about 20°C at both irradiation frequencies of 35 and 94 GHz. The calculated temperatures increase were in good agreement with experiments within the variability of the data. Finally, a 3– D model of the human eye, consisting of realistic anatomical structures, including the cornea, anterior chamber, iris, lens, vitreous humor, and sclera has been considered in [12]. The model takes into account the fluid dynamics of the aqueous humor and predicts a frequency and incident power density dependency to reverse its flow. The model was tested for exposure to 40, 60, 80, and 100 GHz over a range of power density values (10–160 mW/cm2). For instance, at 60 GHz, the study revealed that for incident power density between 10–50 mW/cm2, the direction of the aqueous humor flow is counter clock wise as in the case without exposure. At higher power density of the exposure, specifically from 80 to 160 mW/cm2, the flow direction changes. Moreover, the study showed that the temperature increases linearly with the incident power

208

Appendix A. Millimeter-wave interaction with eye

density of about 0.4°C /10 mW/cm2. The predicted temperature rise for the human eye is lower than the temperature rise computed for the rabbit eye. To sum up, results of those studies demonstrated that the effects of MMW on eyes are dependent on the intensity and duration of exposure. The experimental results show that low- intensity (10 mW/cm2) MMW do not produce any harmful effect on eyes. However, high- intensity exposure results in adverse effects on eye. The ocular damage is induced by the temperature elevation and depends on the frequency, intensity, and duration of exposure.

Figure A.1: Anatomy and structure of the human eye.

209

Appendix A. Bibliography

Bibliography [1] N. Chahat, M. Zhadobov, and R. Sauleau, “Antennas for body centric wireless communications at millimeter wave frequencies,” Progress in compact antennas, Chapter 2, 2014. [2] T. Wu, T. S. Rappaport, and C. Collins, “Safe for generations to come: considerations of safety for millimeter waves in wireless communications,” IEEE Microw. Mag., vol. 16, no. 2, pp. 65–84, 2015. [3] S. Chalfin, J. A. D’Andrea, P. D. Comeau, M. E. Belt, and D. J. Hatcher, “Millimeter wave absorption in the nonhuman primate eye at 35 GHz and 94 GHz,” Health Phys., vol. 83, no. 1, pp. 83–90, Jul. 2002. [4] S. W. Rosenthal, L. Birenbaum, I. T. Kaplan, W. Metlay, W. Z. Snyder, and M. M. Zaret, “Effects of 35 and 107 GHz CW microwaves on the rabbit eye,” USNC-URSI Annu Meet., pp. 110–128, 1976. [5] M. Kojima et al., “Ocular effects of exposure to 40, 75, and 95 GHz millimeter waves,” J. Infrared Millim. Terahertz Waves, vol. 39, no. 9, pp. 912–925, Sep. 2018. [6] H. A. Kues, S. A. D’Anna, R. Osiander, W. R. Green, and J. C. Monahan, “Absence of ocular effects after either single or repeated exposure to 10 mW/cm2 from a 60 GHz CW source,” Bioelectromagnetics, vol. 20, no. 8, pp. 463–473, Dec. 1999. [7] M. Kojima et al., “Acute ocular injuries caused by 60-Ghz millimeter-wave exposure,” Health Phys., vol. 97, no. 3, pp. 212–218, Sep. 2009. [8] A. Papaioannou and T. Samaras, “Numerical model of heat transfer in the rabbit eye exposed to 60-GHz millimeter wave radiation,” IEEE Trans. Biomed. Eng., vol. 58, no. 9, pp. 2582– 2588, Sep. 2011. [9] M. Kojima et al., “Investigation of acute ocular injury threshold by 76 GHz band exposure in rabbits,” in 2011 XXXth URSI General Assembly and Scientific Symposium, 2011, pp. 1–4. [10] K. R. Foster, J. A. D’Andrea, S. Chalfin, and D. J. Hatcher, “Thermal modeling of millimeter wave damage to the primate cornea at 35 GHz and 94 GHz,” Health Phys., vol. 84, no. 6, pp. 764–769, Jun. 2003. [11] A. Karampatzakis and T. Samaras, “Numerical model of heat transfer in the human eye with consideration of fluid dynamics of the aqueous humour,” Phys. Med. Biol., vol. 55, no. 19, pp. 5653–5665, Oct. 2010. [12] A. Karampatzakis and T. Samaras, “Numerical modeling of heat and mass transfer in the human eye under millimeter wave exposure,” Bioelectromagnetics, vol. 34, no. 4, pp. 291– 299, May 2013.

210

Appendix B. Measurement of complex permittivity

Appendix B Measurement of complex permittivity

The complex permittivity of water, culture medium, and agar gel (2.4, Samples under test) was measured at room temperature using an open-ended coaxial probe DAK-1.2E (SPEAG, Zurich, CH) in the 5–65 GHz range. The measurement set-up (Fig. B.1) consists of a: i) vector network analyzer (VNA) operating in the 10 MHz–67 GHz frequency range (Rohde & Schwarz, Munich, DE), ii) PC with a software provided by SPEAG for data acquisition, and iii) single open-ended coaxial dielectric probe (DAK-1.2E) connected to the VNA through a coaxial cable.

Figure B.1: Experimental set-up. The VNA (on background), the PC (on the left), and the open ended coaxial probe mounted on a stand and dipped into the sample under test set on its support (right).

Real and imaginary parts of the complex permittivity of the materials under test are calculated from the reflection coefficient at the probe-material interface [1]. For this reason, the 211

Appendix B. Measurement of complex permittivity measurement accuracy is sensitive to anything that would change the phase or amplitude of the incident and reflected signals along the probe and cable, such as movements of the cable, air gaps or temperature fluctuations. To assure the mechanical stability, the cable was tightly bound to the mechanical support and a flexible mechanical base was used to support the sample and drive it in contact with the probe. Moreover, connections between the probe and the cable, and between the cable and the VNA were kept tight. To assure thermal stability, the sample was left at the desired temperature (i.e., room temperature [RT] of 22 °C) at least 1 h before calibration (section B.1, Calibration procedure) and/or measurement. Before each measure, control of the temperature was done and the probe was carefully rinsed with alcohol to avoid any residual material contaminating the sample under test. During the measurements, the coaxial probe was immersed in the cylindrical plastic flask containing the sample and located on a mechanical support. The complex reflection coefficient measured at the probe end is converted to the complex permittivity of the material under test through the software employed, which uses a method based on the set of formulas published by Ellison and Moreau [2].

B.1 Calibration procedure The calibration normalizes the magnitude and phase changes of the probe and cable so that the reflection coefficient measured by the VNA is normalized to the reference plane at the flange of the probe [1]. Calibration is performed by measuring the reflection coefficient of an open circuit, short circuit, and load. The open calibration is performed by exposing the probe to the air (Fig. B.2a). The short calibration is carried out with a shorting block and metallic strip (e.g., copper, Fig. B.2b). The load calibration requires a material of known dielectric constant. Distilled water has been used in this study.

212

Appendix B. Measurement of complex permittivity

(a) (b)

Figure B.2: (a) Open and (b) short calibration of the open-ended coaxial probe.

B.2 Results Complex permittivity of water, culture medium, and agar gel, at 22 °C, are shown in Figure B.3. Results showed that at 58.4 GHz the complex permittivity of the culture medium is lower than that of water of about 10% (Fig. B.3a). Figure B.3b shows that the complex permittivity of the agar gel is similar to that of water (differences are within the 2%), confirming the fact that the addition of agar in water, at concentration below 4%, does not change the electromagnetic properties of water. The reliability of the experimental procedure used was confirmed by the fact that data obtained for water did not show differences (within the 1%) with values of pure water extracted from the model by Ellison [3] (Fig. B.4).

213

Appendix B. Measurement of complex permittivity

(a)

(b) Figure B.3: Measured complex permittivity of water, culture medium, and agar gel at 22 °C.

214

Appendix B. Measurement of complex permittivity

Figure B.4: Measured complex permittivity of water compared with reference values [3] at 22 °C.

215

Appendix B. Bibliography

Bibliography

[1] Speag, “DAK (Dielectric Assessment Kit) Professional Handbook V1.12”, Schmid & Partner Engineering AG, July 2, 2014. [2] W. J. Ellison and J.-M. Moreau, "Open-ended coaxial probe: model limitations", IEEE Transactions on Instrumentation and Measurement, Vol. 57, no. 9, Sept. 2008. [3] W.J. Ellison, “Permittivity of pure water, at standard atmospheric pressure, over the frequency range 0–25 THz and the temperature range 0–100°C C”, J Phys Chem Ref Data 36(1):1–18, 2007.

216

Appendix C. Measurement of thermal properties

Appendix C Measurement of thermal properties

Specific heat capacity (J/kg∙°C) and thermal conductivity k (W/(m∙°C)) of water and culture medium were measured at the Laboratoire de Thermique et d'Energie de Nantes (LTeN) in the 32–52 °C temperature interval. The experimental protocols are described hereafter.

C.1 Measurement of thermal conductivity Measurements of thermal conductivity were performed using the thermal conductivity analyzer, C-THERM Tci [1] (C-THERM Technologies, New Brunswick, CDN), showed in Figure C.1a. The instrument is based on the Modified Transient Plane Source (MTPS) technique to obtain both thermal conductivity and effusivity (i.e., the square root of the product of the material's thermal conductivity and its volumetric heat capacity), of the material under test. The instrument is made up by a one-sided heat reflecting sensor designed with the heating element which is supported on an insulated backing and surrounded by a guard ring (Fig. C.1b) to allow a one-dimensional heat transfer flow into the sample. Heat is generated when a current is applied simultaneously to the coil and guard ring. A short thermal impulse, typically between 1 to 3 s, results in a rise of the temperature at the interface between the sensor and the sample which induces a voltage drop of the sensor element. Thermal conductivity is inversely proportional to the rate of increase in the sensor voltage drop, or temperature increase (for detailed explanation refer to [2]). The temperature rise will be steeper for lower thermal conductivity materials (e.g., foam) and flatter for higher thermal conductivity materials (e.g., metal).

217

Appendix C. Measurement of thermal properties

(a) (b) Figure C.1: (a) Thermal conductivity analizer C-THERM Tci, sensor, and pc for data acquisition. (b) Zoom on the sensor.

C.1.1. Experimental protocol Calibration of the sensor probe is needed to guarantee correct measurements. In our experiments, calibration was performed by inserting the sensor probe in a glass vessel (100 ml) containing deionized water, at 20 °C, i.e., RT.

Figure. C.2: Calibration of the sensor using deionized water.

For measurements at temperatures different from RT, a thermostat (C Command, Lauda- Koenigshofer, DE) was used to bring and maintain the sample to the desired temperature (Fig. C.3) of 32°C, 42 °C, and 52 °C. The sample under test was set in direct contact with the sensor probe in the same vessel used for calibration avoiding movements of the sensor as well as

218

Appendix C. Measurement of thermal properties formation of bubbles which can affect the quality of the measurements. Sample was left in the water bath until the steady-state temperature was reached (about 1 h) before data acquisition. The increase of the temperature from 32 °C to 52 °C resulted in arising of convective currents as well as bubbles under the probe, slightly reducing the accuracy of the measurement (section C.1.2, Results).

Figure C.3: Measurement of thermal conductivity of the material under test (i.e., culture medium). The glass vessel is embedded in water heated with a thermostat.

C.1.2. Results Results of the measurements of thermal conductivity k of water and culture medium obtained at 20 °C, 32 °C, 42 °C, and 52 °C are presented in the Table C.1. Results are compared with typical values of water of literature [3]. Maximum deviation (less than 2%) of the measured value of water from the literature value was observed at 52 °C.

Table C.1. Thermal conductivity values of water and culture medium in the 20–52 °C temperature interval

k ( ) Literature [3] Measurements Temperature (°C) Water Water Culture medium 20 0.618 0.616 0.612 32 0.62 0.625 0.619 42 0.634 – 0.625 52 0.645 0.636 0.631

219

Appendix C. Measurement of thermal properties

C.2 Measurement of heat capacity

Heat capacity was measured by using the Differential Scanning Calorimetry (DSC) method. Calorimetry is a thermo-analytical technique used to measure the thermal properties of materials by quantifying the heat that is either absorbed or released by the object under a phase transition. The word differential refers to the fact that the technique measures the difference in the heat flow rate, as a function of the temperature, between the sample of interest and a reference (typically air). At constant pressure, heat is referred as enthalpy (H) and it is function of the heat capacity. In this study, a heat flux DSC instrument [4] (Q200, TA Instruments, New Castle, DE) was used for the measurements (Fig. C.4a). The material under test is sealed into a metallic pan (Fig. C4b) through a press (Fig. C.4c). A second pan, typically empty but sealed as well, is used as a reference.

(a) (b) (c) Figure C.4: (a) Differential Scanning Calorimeter (DSC, Q200, TA Instruments, New Castle, DE), (b) pan and lit set, and (c) Tzero Sample Encapsulation Press.

The working principle is schematically represented in Figure C.5a, and briefly described hereafter. The two pans are set into the measurements chamber enclosed in a cylindrical silver heating block (Fig. 5b), which dissipates heat to the specimens by means of a thermoelectric constantan disk. The disk has two raised platforms where the sample and the reference pans are placed. The platforms are connected to the heating block by thin-walled tubes that create thermal resistances between the platforms and the base. Each pan is provided by a thermocouple (TC) positioned in the underside of each platform to measure the difference in temperature during the heating (or cooling) cycle, while another TC is used to measure the temperature of the heating block, and it is embedded in the silver as it serves as controller for the 220

Appendix C. Measurement of thermal properties programmed heating cycle (typically linear). An inert gas is passed through the cell at a constant flow rate of about 40 ml/min. A logistic unit made up by a computer is used to monitor the temperature and regulate the rate at which the temperature of the pans changes. This configuration allows both the pans to be reached by the same amount of heat. However, the heat capacity of the sample determines a temperature difference between the sample and the reference pans, which is measured by the TCs, and is proportional to the heat capacity of the sample. The consequent heat flow is determined by the thermal equivalent of Ohm’s law:

(C.1) where (W/m2) is the heat flow, ΔT (K) is temperature difference between sample and reference, and R (K/W) is the thermal resistance of the thermoelectric disk. The measured heat flow due to the temperature difference between the pans is proportional to the heat capacity of the sample.

(a) (b) Figure C.5: (a) Working principle of the heat flux DSC. (b) Real view of the heating block and the pans.

C 2.1 Experimental Protocol Calibration of the acquisition system was carried out by using two empty pans to acquire the baseline of the instrument, which compensates for the error introduced by the presence of the TC.

221

Appendix C. Measurement of thermal properties

In the post-processing the baseline is subtracted to the measured data to obtain the heat capacity value of the sample under test. Measurements are performed in transient regime. Samples were heated between 10 and 80 °C at a rate of 10 K/m. Post-processing treatment of the data allowed to determine the heat capacity of the samples under test.

C 2.2 Results

Results of specfici heat capacity of water and culture medium at 32 °C and 52 °C are shown in the Table C.2, and compared to values of water of literature [3]. Maximum deviation (< 1%) of the measured value of water from the literature value was observed at 32 °C.

Table C.2. Heat capacity values of water and culture at 32 °C and 52 °C

Literature data [3] Measurements Temperature (°C) Water Water Culture medium 32 4178 4168.9 4067.3 52 4182 4182.7 4080.0

Figure C.6 shows the heat capacity values of water (measured and reference data) and culture medium within the 22–52 °C temperature interval.

222

Appendix C. Measurement of thermal properties

Figure C.6: Heat capacity values of water and culture medium as a function of the temperature.

223

Appendix C. Bibliography

Bibliography

[1] A. OS, TCi Thermal Conductivity Analyzer, C-Therm Thermal Conductivity Instruments.” [Online], Available: https://ctherm.com/products/tci_thermal_conductivity. [2] C-THERM Technologies, “Modified Transient Plane Source (MTPS): Theory and operation”, [Online], Available: http://www.axelproducts.com/downloads/C- Therm_MTPS_Theory_of_Operation.pdf. [3] T. L. Bergman, A. S. Lavine, F. P. Incropera, and D. P. Dewitt, Fundamentals of heat and mass transfer, Seventh Edn. Hoboken, NJ: John Wiley, 2007. [4] T. A. Instruments, “Differential Scanning Calorimeter,” [Online], Available: http://www.tainstruments.com/products/microcalorimetry/differential-scanning-calorimetry/.

224

Appendix D. Measurement of viscosity

Appendix D Measurement of viscosity

Dynamic viscosity η of a fluid is the measure of its internal resistance to flow under the action of an external force i.e., shear stress τ [1]. The latter is defined as the force applied per unit area created by the intermolecular friction exerted when layers of fluid attempt to slide by one another. For Newtonian fluids, such as water, τ is proportional to the shear rate , i.e., the rate at which a progressive shearing deformation is applied to the fluid, with η the proportionality constant. This is defined by the Netwon's law:

(D.1) where is the dynamic viscosity (mPa·s), τ (Pa) the shear stress, and (s-1) the shear rate. Dynamic viscosity of water and culture medium has been measured at the IPR (Institut de Physique de Rennes) by means of a rotational (or shear) rheometer with a cone-plate geometry (MCR 301, Antoon PAAR, Fig. D.1). The resistance to the rotation of the cone produces a torque that is proportional to the shear stress in the fluid. Specifically, the torque measuring system, that connects the drive mechanism to a rotating cone, senses the resistance to rotation caused by the presence of sample fluid between the cone and a stationary flat plate. The resistance to the rotation of the cone produces a torque that is proportional to the shear stress in the fluid. Therefore, viscosity is not measured directly, but is retrieved from the measured interdependence of shear stress and shear rate, and it is function of the torque and the relative angular velocity.

225

Appendix D. Measurement of viscosity

(a) (b) Figure D.1: (a) MCR 301 Rheometer (Antoon PAAR), (b) schematic view of the cone and plate geometry.

Dynamic viscosity of deionized water and culture medium has been measured as a function of the shear rate (Fig. D.2), at 20 °C. Results evidenced that the instrument was not perfectly suitable to measure low viscosity samples (i.e., 0.28–1.3 mPa·s, in the 0–100 °C temperature interval [2]) as it is shown by the fact that the value of η is not constant (deviation of 30% in the interval considered) as a function of as defined by Equation D.1, for Newtonian fluids. However, results obtained qualitatively show that dynamic viscosity of the culture medium is higher than that of water, as confirmed by some studies reported in literature [3], [4] (section 2.4.2, Thermo- physical properties).

226

Appendix D. Measurement of viscosity

Figure D.2: Dynamic viscosity of water and culture medium as a function of the shear rate at 20°C.

227

Appendix D. Bibliography

Bibliography

[1] G.Schramm, ‘A practical approach to rheology and rheometry’, Haake, 1994. [2] T. L. Bergman, A. S. Lavine, F. P. Incropera, and D. P. Dewitt, Fundamentals of heat and mass transfer, Seventh Edn. Hoboken, NJ: John Wiley, 2007. [3] F. Khorshid, “The effect of the medium viscosity on the cells morphology in reaction of cells to topography”, Proc. 2nd Saudi Sci. Conl, Fac. Sci., KAU, pp. 67-98, 2005. [4] E. Fröhlich, G. Bonstingl, A. Höfler, C. Meindl, G. Leitinger, T. R. Pieber, E. Roblegg, “Comparison of two in vitro systems to assess cellular effects of nanoparticles-containing aerosols”, Toxicol In Vitro 27(1):409–17, 2013.

228

About the author

Rosa Orlacchio was born on September 18, 1988 in Sapri, Italy. She received the B.Sc. degree in Clinical Engineering, and the M.Sc. degree in Biomedical Engineering, with honours, from “La Sapienza”, University of Roma, in 2011 and 2014, respectively. She conducted her Master thesis at the Dipartimento di Ingegneria dell'informazione, elettronica e telecomunicazioni (DIET) of “La Sapienza”, University of Roma, in collaboration with the laboratoire de l’Intégration du Matériau au Système (IMS) of Bordeaux. She started working toward the Ph.D degree at the Institut d’Electronique et de Télécommunications de Rennes (IETR), University of Rennes 1, Rennes, France, in October 2015. The research resulted in 3 journal papers (1 in preparation), 3 international conferences, and 1 national conference. Her research interests include the evaluation of the interaction between the millimeter wave and the human body at the cellular and molecular level and the use of the electromagnetic fields for hyperthermia.

229

List of publications

Journal papers

[J–1] R. Orlacchio, M. Zhadobov, S. I. Alekseev, D. Nikolayev, R.Sauleau, Y. Le Page, and Y. Le Dréan, “Millimeter-wave heating in in vitro studies: effect of convection in continuous and pulse-modulated regimes”, Bioelectromagnetics, accepted on September 2019. [J– 2] R. Orlacchio, Y. Le Page, and Y. Le Dréan, R. Le Guével, R. Sauleau, S.I. Alekseev, and M. Zhadobov,”Millimeter-wave pulsed heating in vitro: cell mortality and heat shock response ”, Sc.Report, accepted on September 2019.

International conference papers

[IC–1] R. Orlacchio, Y. Le Page, and Y. Le Dréan, R. Le Guével, R. Sauleau, S.I. Alekseev, and M. Zhadobov, “Heat shock response of melanoma cells induced by continuous and pulsed millimeter-wave heating”, in BioEM 2019, Montpellier, France, 23–28 June 2019. [IC–2] R. Orlacchio, M. Zhadobov, S. I. Alekseev, D. Nikolayev, R.Sauleau, Y. Le Page, D. Habauzit, and Y. Le Dréan, “Thermal convection in in vitro sample exposed to millimeter wave in continuous and pulse-modulated regimes”, in BioEM 2019, Montpellier, France, 23–28 June 2019. [IC–3] R. Orlacchio, M. Zhadobov, S. I. Alekseev, R.Sauleau, Y. Le Page, and Y. Le Dréan, “Local thermal dosimetry applied in in vitro studies at millimeter wave”, in BioEM 2018, Piran- Portoroz, Slovenia, 24–29 June 2018.

National conference papers

[NC–1] R. Orlacchio, M. Zhadobov, S. I. Alekseev, R.Sauleau, Y. Le Page, and Y. Le Dréan, “In vitro millimeter-wave heating: role of thermal convection”, in Journée scientifique Effets Biologiques et Sanitaires des Rayonnements Non-Ionisants, 2 October 2018, in Montpellier, France.

230

Acknowledgment

The few words scribbled below will never be sufficient to express my deep gratitude to all the people who made great this journey. First, I would like to thank my supervisor, Professor Ronan Sauleau who accepted my candidacy allowing me to join the IETR. Moreover, I wish to express my real gratitude to my co-supervisor Dr Maxim Zhadobov who supported my research day-by-day teaching me the rigour of science. I would like to express my sincere gratitude to Professor Theodoros Samaras and Dr. Delia Arnaud-Cormos for accepting to review my manuscript. Besides, I would like to thank my dear Professor Micaela Liberti for being a member of my jury committee, and to have firstly introduced me into the world of bioelectromagnetics during my Master. I would like to sincerely present my profound gratitude to Dr. Yann Le Page and Dr Yves Le Dréan who warmly welcomed me into their team of biologists. They patiently taught me the basis of molecular biology, inspiring my future research. I would like to express my deep gratitude to Dr. Stan Alekseev. His deep knowledge of MMW, his constant availability, our Skype talk at any time to solve my doubts made possible the success of this work. In addition, I'd like to thank the Sp@rte team of the IGDR (Institut de Génétique et Développement de Rennes) who shared their cell culture room, and particularly Audrey. I wish to thank all the members of the IETR: the team of the technicians and above all the secretaries Martine, Nathalie, and Noëlle for all their help and smiles. Moreover, I wish to thank Professors Colombel and Himdi for their happiness and support through these years. I would like to thank all my officemates Evgeny, Oumar, Gill, Seydouba, Yann, and Nicolas, as well as all the boys Nicolas, Sdran, Jorge, Antoine, Zain, François, Tuyen, and Denys (and all our fruitful discussions and help with simulations) for making shorter the long working day, and specially Alvin, Michelino, Andrea, and Nicola for all the pizzas and parties we had. I wish to thank my little girls, my compañer as de merienda (and not only) Laura and Fatima. Thank you for our talks, biscuits, teas, sushi, and complaints. I'm truly missing you here. I want to thank all my flatmates, the double Eric and Thibault, who taught me French and made pleasant my stay in Rennes. I'd like to warmly thank my dears Fabrizio, Fausta, Ida, Riccardo, and Steph for all the moments we shared. You always made my days better and exciting, constantly providing me support, laughs, love, food, and beers. Special thanks to Kaledino and Francibeth, and their true happiness. I wish to heartily thank Francesco and Yannis, who guided me throughout my thesis making IETR feel home. Francesco, his class, wisdom, and friendship have been precious to me, and always will be. Yannis, the Greek friend everyone should have, who always solved all my technical 231

troubles (and still does), who made easier my life in Rennes and my long nights in the lab, and even longer in the city center. Thank you Ioannis, for all the beers, the Saturdays at L' Amaryllis, the Gipsy King songs, the night talks, the dancing parties, the laughs and cries, the hopes, joys and sorrows we shared. I want to thank Andrea for all his Spritz and craziness. I want to acknowledge Benjamin, Wendy, and Joris for their kindness, and Maria for all our sparkling, supportive, and stimulating conversations. Least, but not last, I wish to thank my Italian friends, spread all over the world, but so close to my heart. I wish to thank my mum, dad, sister, and grandma for their support. Thank you to have always believed in me, being proud of each of my little successes, constantly. I want to profoundly thank my Uriolino, who followed me in this adventure with all the love and patience of the world. To more to come! Sincerely yours,

Rosa

232

Titre Électromagnétisme pour le biomédical : étude des changements induits au niveau cellulaire par des impulsions thermiques générées par les ondes millimétriques

Mots clés : Ondes millimétriques, convection thermique, expériences in vitro, chauffage induit par impulsions, réponse aux chocs thermiques, destruction non invasive des cellules.

Résumé : La partie inférieure de la bande de les vitro typiques lors du chauffage par CW et PWM est ondes millimétriques (OMM, c'est-à-dire 20–100 étudié expérimentalement. Deuxièmement, la GHz) constitue une alternative attrayante pour le réponse au choc thermique, médiée par la traitement thermique non invasif du mélanome. Le phosphorylation d'une protéine de choc thermique chauffage pulsé induit électromagnétiquement peut (HSP27) et l'activation de Caspase-3, indicateur de entraîner des dommages plus importants dans les l'apoptose cellulaire, est quantifiée pour surveiller la cellules par rapport au chauffage continu réponse biologique en utilisant une approche traditionnel. Dans ce travail, nous étudions les expérimentale basée sur la microscopie à modifications induites au niveau cellulaire dans les fluorescence. Deux durées d'impulsion (1.5 s et 6 s) cellules de mélanome à la suite d'une exposition au sont considérées. Nos résultats démontrent que les chauffage induit par en onde continue (CW) et ou impulsions thermiques sont capables d'induire une en régime modulé (PWM) avec avec la même réponse cellulaire plus forte dans les cellules de élévation de température moyenne, à 58.4 GHz. mélanome à la fois en termes de choc thermique et Premièrement, l’impact de la convection thermique de mortalité cellulaire par rapport à celle induite en sur la dynamique de la température dans des CW. Plus la durée de l'impulsion est courte, plus la modèles représentant des conditions d’exposition in réponse cellulaire est grande.

Title: Millimeter waves for biomedical electromagnetics: study of changes induced at the cellular level by pulsed electromagnetically-induced heating

Key words : Millimeter waves, thermal convection, in vitro experiments, pulsed-induced heating, heat shock response, non-invasive destruction of cells.

Abstract : The lower part of the millimeter wave typical in vitro exposure conditions during CW and (MMW) band (i.e., 20–100 GHz) is an attractive PW-induced heating is experimentally investigated. alternative for non-invasive thermal treatment of Second, the heat shock response, mediated by melanoma. Besides, pulsed electromagnetically- phosphorylation of a small heat shock protein induced heating can lead to stronger damage in (HSP27) and activation of Caspase-3, indicator of cells compared to traditional continuous heating. In cellular apoptosis, are quantified to monitor the in-vitro experiments, continuous-wave (CW) or biological response using an experimental pulsed-wave (PW) amplitude-modulated MMW can approach based on fluorescence microscopy. Two be efficiently used to locally heat cell monolayers pulse durations (1.5 s and 6 s) are considered. Our with a typical thickness ranging between 3 µm and results demonstrate that thermal pulses are able to 10 µm. In this work we investigate the modifications induce a stronger cellular response in melanoma induced at the cellular level in melanoma cells cells both in terms of heat shock and cellular following exposure to CW and PW MMW-induced mortality compared to the one induced by CW. The heating with the same average temperature rise, at shorter the pulse duration, the greater the cellular 58.4 GHz. First, the impact of thermal convection response. on temperature dynamics in models representing