Research Collection

Doctoral Thesis

Therapeutic Cell Engineering: Electrogenetics, Bioelectronic Implant Design, and Rewiring of Intracellular Signaling Pathways Using dCas9

Author(s): Krawczyk, Krzysztof

Publication Date: 2020

Permanent Link: https://doi.org/10.3929/ethz-b-000447722

Rights / License: In Copyright - Non-Commercial Use Permitted

This page was generated automatically upon download from the ETH Zurich Research Collection. For more information please consult the Terms of use.

ETH Library DISS. ETH NO. 26692

Therapeutic Cell Engineering: Electrogenetics, Bioelectronic Implant Design, and Rewiring of Intracellular Signaling Pathways Using dCas9

A thesis submitted to attain the degree of

DOCTOR OF SCIENCES of ETH ZURICH

(Dr. sc. ETH Zurich)

presented by

Krzysztof KRAWCZYK

Magister biotechnologii, Uniwersytet Jagielloński

Born on 27.11.1986

Citizen of

Poland

Accepted on the recommendation of Prof. Dr. Martin Fussenegger / Examiner Prof. Dr. János Vörös / Co-Examiner Prof. Dr. Petra S. Dittrich / Co-Examiner Prof. Dr. Randall Platt / Co-Examiner

2020

Page 2 of 200

Table of Contents

Summary ...... 5 Zusammenfassung ...... 7 Introduction ...... 9 Synthetic biology and bioengineering ...... 9 Therapeutic cell engineering ...... 10 Chimeric antigen receptor T-cells...... 11 Designer cells ...... 12 Regulation of cell behaviour ...... 14 Transcriptional regulation ...... 14 Post-transcriptional regulation ...... 14 Post-translational regulation ...... 15 Timescale ...... 17 Bio-electronic interfaces for therapeutic cell engineering ...... 18 Cell encapsulation ...... 20 Microencapsulation ...... 21 Macroencapsulation ...... 21 Cellular therapies for type 1 diabetes ...... 24 Contribution of this work ...... 25 Contribution of CHAPTER I: “Electrogenetic cellular insulin release for real-time glycemic control in type-1 diabetic mice.” ...... 25 Contribution of CHAPTER II: “Rewiring of endogenous signalling pathways to genomic targets for therapeutic cell reprogramming.” ...... 26 CHAPTER I: Electrogenetic cellular insulin release for real-time glycemic control in type-1 diabetic mice ...... 27 Results ...... 31 Discussion ...... 46 Supplementary Materials ...... 50 Materials and Methods ...... 51 Supplementary Figures ...... 60 Supplementary Tables ...... 87 CHAPTER II: Rewiring of endogenous signalling pathways to genomic targets for therapeutic cell reprogramming ...... 101 Abstract ...... 101

Page 3 of 200

Introduction ...... 102 Results ...... 105 Discussion ...... 114 Methods ...... 117 Supplementary Information ...... 121 Discussion ...... 169 Electrogenetic cellular insulin release for real-time glycemic control in type-1 diabetic mice ..... 169 Rewiring of endogenous signalling pathways to genomic targets for therapeutic cell reprogramming ...... 173 References ...... 175 Acknowledgements ...... 199 Curriculum Vitae ...... Error! Bookmark not defined.

Page 4 of 200

Summary

Cell therapies utilize functions of the entire living cell to fight diseases. The first chimeric antigen receptor T-cells are already approved for cancer treatment and many others are being developed. Furthermore, multiple proof-of-concept animal studies showed promising results in the treatment of metabolic diseases using designer cells. To face the challenges of their clinical translation, new purpose-driven methods to regulate the behavior of engineered cells are required.

Electronic devices can collect numerous diagnostic or biometric data. Combining them with therapeutic cells to create hybrid, bioelectronic systems could leverage strengths of both components. To achieve this, it would be essential to create an electrogenetic interface that translates the information collected by electronic devices to a format which engineered cells are able to interpret. CHAPTER I of this thesis presents the first direct interface between therapeutic cells and electronic devices. Electrical stimulation opens L-type voltage gated calcium channels and causes calcium influx, which can be linked to either transgene expression, or fast vesicular secretion. The resulting engineered human cell line Electroβ is capable of rapid insulin secretion in the timescale of minutes, bypassing the transcriptional delay typical for current synthetic systems. A wireless-powered subcutaneous implant encapsulating Electroβ cells can provide real-time control of glycaemia in a type I diabetes mouse model. Such electrogenetic devices offer new opportunities for advanced healthcare in the future.

Reprogramming of cellular behavior typically focuses on achieving an input-specific reaction. Current designs that rely on engineered receptors are limited to single inputs, and often suffer from high leakiness and low fold induction. CHAPTER II of this thesis focuses on a new method of rewiring of endogenous signaling pathways to alternative genomic targets, to upregulate expression of genes important for therapeutic purposes. It introduces Generalized Engineered Activation Regulators (GEARs) that consist of the MS2 bacteriophage coat protein fused to regulatory or transactivation domains. GEARs are driven by catalytically inactive Cas9 (dCas9), and can hijack intracellular signaling dependent on NFAT, NFκB, SMAD2 and Elk1. Because of being pathway-specific, they can integrate and process multiple input signals. GEARs enable a membrane depolarization-induced activation of insulin production in β-mimetic cells, interleukin 12 expression in activated immortalized T-cells (Jurkat), interleukin 12 production in response to the immunomodulatory cytokines TGFβ and TNFα in HEK293T cells, as well as a simultaneous activation of two genes. Engineered cells with the

Page 5 of 200 ability to reinterpret their behavioral programs have potential for applications in immunotherapy and in the treatment of metabolic diseases.

Page 6 of 200

Zusammenfassung

Zelltherapien nutzen die Funktionen der gesamten lebenden Zellen zur Bekämpfung von Krankheiten. Die ersten Chimären Antigenrezeptor-T-Zellen sind bereits für die Krebsbehandlung zugelassen, und viele weitere befinden sich im Entwicklungsstadium. Darüber hinaus zeigten mehrere Proof-of-Concept-Tierstudien vielversprechende Ergebnisse bei der Behandlung von Stoffwechselkrankheiten mit Designerzellen. Um sich den Herausforderungen ihrer klinischen Umsetzung zu stellen, sind neue zweckgerichtete Methoden zur Regulierung des Verhaltens von manipulierten Zellen erforderlich.

Elektronische Geräte können zahlreiche diagnostische oder biometrische Daten sammeln. Die Kombination mit therapeutischen Zellen zur Schaffung hybrider, bioelektronischer Systeme könnte die Stärken beider Komponenten nutzen. Eine elektrogenetische Schnittstelle, die die von elektronischen Geräten gesammelten Informationen in eine Form übersetzen würde, welche von entwickelt Zellen interpretiert werden könnte, wäre ein wesentliches Bindeglied. KAPITEL I dieser Arbeit stellt die erste direkte Schnittstelle zwischen therapeutischen Zellen und elektronischen Geräten dar. Die elektrische Stimulation öffnet spannungsgesteuerte Kalziumkanäle des L-Typs und verursacht einen Kalziumeinstrom, der entweder mit der Expression von Transgenen oder mit einer schnellen vesikulären Sekretion verbunden sein kann. Die daraus resultierende künstlich hergestellte menschliche Zelllinie

Electroβ ist in der Lage, Insulin innerhalb von Minuten schnell zu sekretieren und damit die für die derzeitigen synthetischen Systeme typische Transkriptionsverzögerung zu umgehen. Ein drahtlos betriebenes subkutanes Implantat, das Electroβ Zellen einkapselt, ermöglicht die Echtzeitkontrolle von Glykämie im Mausmodell für Typ-I-Diabetes. Elektrogenetische Geräte bieten neue Möglichkeiten für die fortgeschrittene Gesundheitsversorgung der Zukunft.

Die Neuprogrammierung des Zellverhaltens konzentriert sich in der Regel auf das Erreichen einer input-spezifischen Reaktion. Die derzeitigen Designs, die sich auf technisch entwickelte Rezeptoren stützen, sind auf einzelne Inputs beschränkt und leiden oft unter hohen basalen und geringen maximalen Aktivierungsleveln. KAPITEL II dieser Arbeit konzentriert sich auf die Neuverkabelung endogener Signalwege zu alternativen gnomischen Zielen, um die Expression von Genen, die für therapeutische Zwecke wichtig sind, hochzuregulieren. Es werden Generalized Engineered Activation Regulators (GEARs) vorgestellt, die aus dem Hüllprotein des MS2-Bakteriophagen bestehen, das mit regulatorischen oder Transaktivierungsdomänen fusioniert ist. GEARs werden durch katalytisch inaktives Cas9 (dCas9) angetrieben und können in Abhängigkeit von NFAT, NFκB, SMAD2 und Elk1 Page 7 of 200 intrazelluläre Signale umleiten. Da sie pfadspezifisch sind, können sie mehrere Eingangssignale integrieren. GEARs ermöglichten die Membrandepolarisations-induzierte Aktivierung der Insulinproduktion in β mimetischen Zellen, die Interleukin-12-Expression in aktivierten immortalisierten T-Zellen (Jurkat), die Interleukin-12-Produktion als Reaktion auf die immunmodulatorischen Zytokine TGFβ und TNFα in HEK293T-Zellen sowie die gleichzeitige Aktivierung zweier Gene. Umgebaute Zellen mit der Fähigkeit, ihre Verhaltensprogramme neu zu interpretieren, könnten Anwendungen in der Immuntherapie und bei der Behandlung von Stoffwechselkrankheiten finden.

Page 8 of 200

Introduction

The dawn of biotechnology can be dated back to 12 000 BC, when the first livestock was domesticated. The desire to improve nature by elevating useful features in animals led to a slow, artificial selection process that lasts until today1. Scientific progress in the 20th century transformed biotechnology into a rapidly growing, multi-purpose oriented field with an immense impact on medical and pharmaceutical sciences2. The development of genetic manipulation techniques, initially limited to single gene overexpression, now enables designing sophisticated biological systems. Hence, synthetic biology and bioengineering became an inspiration for creating new treatment strategies and improve human life.

Synthetic biology and bioengineering

Although it is difficult to clearly distinguish synthetic biology and bioengineering, it is reasonable to understand bioengineering in broader terms. Synthetic biology, defined as “designing and constructing biological modules, biological systems, and biological machines or, re-design of existing biological systems for useful purposes”3, is a part of biology, while bioengineering is a truly interdisciplinary field, which applies synthetic and analytical principles of engineering to biomedical technologies4. Synthetic biology emerged from genetic engineering, as complexity of the designed systems increased. Development of new methods and a rise in the amount of available data, as well as a greater capability to analyze biological systems allowed not only for production of single proteins, but for the engineering of whole metabolic or regulatory pathways. Synthetic biology-inspired strategies find application in designing biosensors5, industrial production of biological compounds6, plant biotechnology7, therapeutic cell engineering8-11, programmable cell differentiation12,13, and even in basic research14. Recently, first cellular therapies have been approved for clinical use15 and many others have been applied to combat diseases in multiple animal models11. Combining living cells with electronic devices aimed at joining advantages of both and lead to engineering proof- of-concept biomedical implants16,17 – an example of an overlap between synthetic biology and bioengineering.

Page 9 of 200

Therapeutic cell engineering

Formally, the first successful attempt to utilize living human cells for a therapeutic purpose can be dated back to 1968, when the first successful bone marrow transplantation took place18. Despite of being a black-box approach at that time, cell and organ transplants started to gain engineering aspects. The possibility of using rationally reprogrammed cells was investigated in multiple therapeutic areas, mainly cancer treatment, handling chronic diseases and organ regeneration19.

The main application field for therapeutic cells are situations in which the therapeutic effect requires, or is significantly improved, by the whole cell function. Immune cells, which are naturally designed to fight pathogens and eliminate malfunctioning cells, can be infused into a patient to either directly kill cancerous or virus-infected cells, or modulate the host’s immune response20-22. Cytotoxic T-cells were extensively tested for treating cancer and viral infections23,24, since it was discovered that allogenic immune cells obtained during bone marrow transplant have an anticancer effect25. Since then, bone marrow transplants became the method of choice for blood cancer treatment and, furthermore, protocols for ex vivo immune cells activation have been tested for solid tumors26. Despite of promising results, these methods are strongly limited by the number of available cells, high patient-dependent variability, or the time-consuming manufacturing process24. The whole cell function can be also achieved by introducing new functions to cells that were not natively designed to serve them. Such synthetic systems were investigated as next-generation therapies for chronic diseases8-11. The functional principle of therapeutic cell engineering is presented in Figure 1 below.

Page 10 of 200

Figure 1. The functional principle of therapeutic cell engineering. Engineered programs can regulate the therapeutic cell function by sensing and processing external signals, cell-bound markers, or soluble systemic disease markers. Synthetic gene networks, as well as intracellular signaling pathways can lead to production of a therapeutic protein, secretion of a pre-produced protein, or modify the behavior of the entire cell. Theranostic systems rely on a negative feedback loop to self-regulate and adjust the level of the therapeutic protein production to the level of the disease marker.

Chimeric antigen receptor T-cells

Redirecting T-cell response with engineered receptors was proposed as a solution to bypass their low efficiency in killing autologous cells – a major limitation of adoptive cell anti- cancer therapies. The invention of chimeric antigen receptors (CAR), and their incorporation into native signaling pathways of T-cells led to development of clinically-feasible therapeutic methods15,27,28, which recently received clinical approvals for lymphoma treatement15. Chimeric antigen receptor T-cells (CAR-T) are patient-derived T-cells that were reprogrammed to recognize a cancer-related antigen and kill the cancer cell29. The functional principle of constructing CARs involves fusing an extracellular antigen-binding domain with an intracellular domain capable of activating cytotoxic response of a T-cell. Although such approach gives promising outcome in blood cancer treatment, solid tumors remain a challenge due to their immunosuppressive microenvironment and physical barriers, which have to be surpassed by therapeutic cells30. That issue has been addressed by the investigation of multiple variants of next-generation CAR-T cells. The production of immunostimulatory compounds, Page 11 of 200 like interleukin 2 (IL-2) or interleukin 12 (IL-12), either expressed upon CAR activation, or in response to tumor environment-related protein, have been shown to improve tumor infiltration and killing cancer cells30. Such modified CAR-T cells are often referred to as “armored CAR- T”30. Alternative methods rely on downregulating response to immunosuppressive cytokines. Overexpressing a dominant negative mutant of transforming growth factor β (TGFβ)31-33 or reverting response to interleukin 4 (IL-4) by overexpressing a chimeric receptor with IL-4 extracellular domain and immunostimulatory interleukin 7 (IL-7) domain34 prevent deactivating CAR-T in proximity of tumors. Additionally, to increase the extravasation, they can be engineered to recognize components of extracellular matrix or vascular endothelial growth factor receptor (VEGFR), usually overexpressed on blood vessels epithelium on tumor sites35.

CAR-T cell therapy is related to various adverse effects, of which cytokine release syndrome (CRS) is considered the most serious. CRS is a life threatening condition, related mostly to immunotherapy of blood cancers, but can be observed with solid tumors as well30, caused by a rapid release of large amounts of cytokines into the blood from immune cells36. This and other types of side effects, like on-target/off-tumor toxicity30, can be partially counteracted by optimizing the dose and time of cells administration; however, they remain a major challenge in antitumor cell engineering. The design strategies aimed at eliminating these adverse effects are focused on synthetic biology-based logic gates. Examples include co- expressing two receptors37, designing a tandem CAR38,39, or adding an inhibitory receptor (iCAR) to suppress the unspecific response40.

Designer cells

Many chronic diseases develop and progress asymptomatically. Many require continuous monitoring and the repetitive administration of a drug. Consequently, the clinical outcome is strongly dependent on patient’s self-discipline and awareness. Meanwhile, non- adherence to medication negatively affects nearly 50% of patients and accounts for several percent of hospitalization cases41-43. Cellular theranostic (diagnostic + therapeutic) devices can constantly monitor the level of a disease marker and deliver medication when required, even before the symptoms occur9, as demonstrated in the case of psoriasis44. Self-regulating systems can also act as preventive guards, providing information about an occurrence of a pathological condition with a readable output, such as a biomedical tattoo45. While most theranostic cellular devices rely on protein as an input signal, some conditions can be diagnosed by change in other parameters, like high blood glucose level. Due

Page 12 of 200 to the prevalence of the diseases and defined way of treatment, an extensive effort has been made to develop designer cell implants against type I diabetes8,16,46-49. In a theranostic approach, the functional principle of pancreatic β-cells was recreated in human endothelial kidney cells (HEK293T). Membrane depolarization-based glucose sensing was achieved by overexpressing the L-type voltage-gated calcium channel and coupling high glucose-induced calcium influx to transgenic insulin expression46. Meanwhile, blood glycaemia can be measured fast and precisely with diverse biosensors that are currently available on the market. Hence, a system in which an external readout provides information when to trigger a cellular response offers an alternative strategy, which was applied for instance in light-inducible16 or magnetic field- inducible48,49 systems. Proteins and peptides, which are therapeutic agents produced by engineered cells, can be manufactured in a bioprocess. Unfortunately, the final products are sensitive to storage conditions and undergo degradation and aggregation, limiting their shelf-life50.life50. An additional advantage of therapeutic cells over traditional therapies is their ability for on-site production of a therapeutic protein, which abolishes the problem of storage-related inactivation and repetitive injections.

Page 13 of 200

Regulation of cell behaviour

The functionality of engineered cells depends on robust, accurate and solution-oriented regulation system. The need for various types of molecular switches is reflected in numerous approaches that have been developed up to date in order to act at all possible control levels. This chapter focuses on three main types of regulation: transcriptional, posttranscriptional, and posttranslational.

Transcriptional regulation

Transcription factors (TF) regulate gene expression in eukaryotic cells by forming multiprotein complexes with co-modulators that bind DNA and either activate or inhibit transcription51. Despite of the higher complexity of the described process comparing to prokaryotes, first transcriptionally regulated genetic circuits in mammalian cells were constructed by adapting solutions known from bacteria52,53. DNA binding domains from bacterial TFs, together with their regulatory domain, allow for docking TFs to the operator sequences. Fusing them to transactivators (VP16, p65TA, HSFTA), inhibitory domains (KRAB) or nuclear receptors, create synthetic mammalian transcription factors (smTF) which can respond to antibiotics, hormones, metabolites, food additives, or cosmetics54-56 (Fig. 2a). Other systems rely on natural signaling pathways, which evolved for precise and tunable regulation of gene expression. They were utilized to couple the signal from a natural receptor to a transgene expression via calcium signaling46,57, cyclic AMP (cAMP)58, cyclic GMP (cGMP)16,17,59, Nuclear Factor κB (NFκB)44, or heat shock proteins60 (Fig. 2b). These approaches, however, are limited to the regulation of transgene expression. Regulating expression of endogenous genes requires the usage of proteins that can be programmed to recognize desired DNA sequences, like zinc finger (ZF) domains, transcription activator-like effectors (TALEs) or catalytically dead clustered regulatory interspaced short palindromic repeats (CRISPR)- associated protein 9 (dCas9)61. Due to its modularity and simplicity, dCas9-based designs became the most widely applied for creating inducible systems triggered by small molecules62, hormones63,64, or light65,66.

Post-transcriptional regulation

Posttranscriptional regulation can be achieved through modulating a translation process, RNA stability or mRNA splicing10. Most of these strategies involve RNA aptamers - folded RNA structures capable of binding ligands. Incorporating a protein-binding aptamer to the 5’- untranslated region of mRNA allows for binding an inhibitor, which blocks the translation

Page 14 of 200 process67 (Fig. 2c). Similar elements can be integrated into regulatory forms of RNA: short hairpin RNA (shRNA)68 or to microRNA69. Stability of mRNA can be modulated by catalytically active aptamers - self-cleaving aptazymes70. Ligand binding modulates the catalytic activity of the aptazyme, which causes either mRNA stabilization, or degradation (Fig. 2d). In another setting, a tetracycline-binding riboswitch is used to promote exon skipping and alternate splicing71.

An alternative approach to posttranscriptional regulation was used in designing a caffeine-inducible global translational downregulation. That system relies on protein kinase R, which has an ability to switch off most cellular translation processes (Fig. 2e). Caffeine-induced kinase R dimerization downregulates classical translation, while the transgene that is placed downstream of an internal ribosomal entry site is resistant to that downregulation72.

Post-translational regulation

Regulation of a protein level can occur by its controlled degradation. Tagging a transgenic protein with a degradation signal can be used to modify its half-life and change magnitudes of protein expression73 (Fig. 2f). Inducible protein degradation was achieved with systems triggered by a plant hormone74, or a small molecule75-80. In the first case, auxin- inducible protein degradation was achieved by transferring components of a plant degradation system into mammalian cells. Small molecule-inducible degradation system was designed by tagging proteins with an engineered dehalogenase (HaloTag). Upon binding a small molecule, the HaloTag marks the fusion protein as unfolded, which triggers its degradation75. Similarly, a protein fused to a destabilizing domain can be stabilized by a ligand addition, while w protein fused to a stabilizing domain can by destabilized76-80.

Besides modulating protein stability, posttranslational regulation can also be achieved by inducible protein secretion (Fig. 2g) – a process widely used by natural hormone-secreting cells. For example, an engineered variant of the human FK506-binding protein FKBP12 (FM) forms aggregates in endoplasmic reticulum. Upon small molecule ligand addition, the aggregate is dissolved and protein secretion occurs. Fusing FM to recombinant insulin via a furin cleavage site allows for its inducible secretion from engineered cells81.

Page 15 of 200

Figure 2. Schematic representation of selected methods used in synthetic biology to regulate cell behavior. (a) Engineered transcription factors activate gene expression either by ligand-induced DNA binding, or by ligand-induced recruitment of the transcription machinery. (b) Ligand binding by a membrane receptor activates native endogenous signaling pathways and upregulates transcription. (c) An aptamer incorporated into the 5’ Untranslated Region (5’UTR) binds an inhibitory protein that blocks translation. (d) A self-cleaving aptazyme incorporated into 3’ Untranslated Region (3’UTR) decreases half-life of RNA by promoting its degradation. Ligand binding by the aptazyme inhibits self-cleaving and stabilizes RNA. (e) The activation of Protein Kinase R (PKR) blocks the translation process in the cell. (f) A degradation tag fused to a protein modifies the protein half-life. An inducible degradation can be achieved by ligand-induced activation of protein degradation. (g) Protein molecules aggregated in the endoplasmic reticulum can be released upon addition of a small molecule ligand.

Page 16 of 200

Timescale

A time delay between cell activation and secretion of a newly produced protein is a major limitation of transcriptionally regulated synthetic systems, especially for time-sensitive applications, like regulation of blood glucose level. When a signaling molecule binds to its extracellular receptor, the intracellular signaling pathway is triggered to initiate transcription. The speed of that process depends on the stimulation strength and a signaling pathway involved. For calmodulin / calcinerurin / NFAT pathway it takes at least 5 – 10 minutes until the transcription factor translocates to the nucleus82. Mammalian cells transcribe RNA at the rate of 10-100 nucleotides per second, which gives 10 minutes for a ‘typical” gene (10 000 base pairs)83. Pre-mRNA splicing takes between 0.4 to 7.5 minutes and nuclear export was estimated for approximately 4 minutes84. Translation occurs at a speed of approximately 10 amino acids per second, which gives 1 minute per protein83. Posttranslational protein processing and secretion is a more complicated process and its time depends on many factors. In kinetic analysis for a small, soluble and fast-folding C-terminal domain of the Semliki Forest virus capsid protein, the authors observed that first molecules were secreted after 15 minutes after production85. However, they noticed that it is the fastest time recorded for mammalian cells and that the total secretion time of a translated pool took more than 2 hours. Overall, the typical time for protein secretion after initiation of the transcription can be estimated at least 30 minutes for small non-complex proteins.

In contrast, the secretion of pre-produced insulin-containing vesicles from a pancreatic β-cell takes only a few seconds after an increase in cytosolic ATP level86 and a few minutes after increase in extracellular glucose level87.

Page 17 of 200

Bio-electronic interfaces for therapeutic cell engineering

Communicating an electronic device with actuating therapeutic cells requires translating an electrical signal into a form which can be recognized by living cells. Typically, such bioavailable signals include proteins, peptides, or small molecules. Fortunately, cellular response can be also triggered by physical stimuli, like heat or light, which are generally more compatible with electronic devices. In the history of mammalian synthetic biology, heat- inducible systems based on heat shock factor promoters-driven transgene expression were the first ones in that relied on a physical stimulus88. Since then, that concept was further developed and heat was transmitted by magnetic fields (magnetogenetics) or radio waves (radiogenetics) 49,89-91 to facilitate in vivo applications . However, the most widely applied physical stimulus was light16,17,47,92-95. The abundance of light-sensing proteins in nature and their diverse functions established a foundation for optogenetics47,96,97 and for the development of first bio-electronic implants encapsulating therapeutic cells16,17. A separate class of bio-electronic interfaces consists of systems that rely on activating cells by products of an electrochemical reaction. Although that approach was tested in mammalian cells95, it was more promising in bacteria98. Examples of physical stimulus-controlled systems are listed in Table 1.

Table 1 Physical stimulus-controlled systems for biomedical applications

Ultimate Direct Element System Application Reference inducer inducer

Human Blue light-inducible insulin Diabetes Ye et al. Blue light Blue light melanopsin expression treatment 201157

Synthetic Blue light inducible transcription homodimerization of a synthetic Diabetes Wang et Blue light Blue light factor transcription factor initiates treatment al. 201299 [Gal4(65)- transgene transcription VVD]

Modified Blue light-inducible synthetic Beggiatoa sp. guanylate cyclase producing a Erectile Kim et al. Blue light Blue light PS BlaC-derived cyclic guanosine dysfunction 201593 guanylate monophosphate (cGMP) leads treatment cyclase to penile erection.

Optogenetic control of a Kushibiki Blue light Blue light Channelrhodops Diabetes calcium channel causes insulin et al. (laser) (laser) in-2 (ChR2) treatment secretion 201547

Page 18 of 200

Ultimate Direct Element System Application Reference inducer inducer

Low- Radio-wave heating of frequency Genetically genetically encoded radio Heat encoded ferritin Diabetes Stanley et nanoparticles causes calcium waves or (?100,101) nanoparticles, treatment al. 201549 influx via TRPV1 leading to magnetic TRPV1 insulin expression field

Stress-inducible Magnetic Magnetic nanoparticles heating Antitumor Ito et al. Heat promoter field by magnetic field therapy 2001 102 gadd153

Radio-wave heating of Genetically Behavioral genetically encoded Magnetic Heat encoded ferritin control of Wheeler et nanoparticles causes calcium field (?100,101) nanoparticles, zebrafish al. 2016103 influx via TRPV4 leading to TRPV4 and mice neurons excitation

Artificial heat Magnetic nanoparticles shock-inducible encapsulated together with Magnetic Proof of Ortner et Heat promoter, therapeutic cells expressing field concept al. 201290 magnetic transgene under control of heat nanoparticles shock-responsive promoter

Biotinylated magnetic nanoparticles Magnetic-field heating of cell and engineered Behavioral Magnetic membrane-bound nanoparticles Huang et Heat membrane control of a field opened heat sensitive TRPV1 al. 201091 protein marker worm channel with a biotin acceptor peptide, TRPV1

Near infrared light-activated Bacterial bacterial diguanylate cyclase Mental state- Near Near diguanylate (DGCL) produces cyclic controlled Folcher et infrared infrared cyclase diguanosine monophosphate (c- transgene al. 201417 light light (DGCL), di-GMP), which triggers the expression STING Stimulator of Interferon Genes (STING)-dependent induction of a synthetic promoter

Bacterial Near infrared light-activated Smartphone- Near Near diguanylate bacterial diguanylate cyclase controlled Shao et al. infrared infrared cyclase (DGCL) produces cyclic insulin 201716 light light (DGCL), diguanosine monophosphate (c- expression STING di-GMP), which triggers the Stimulator of Interferon Genes

Page 19 of 200

Ultimate Direct Element System Application Reference inducer inducer

(STING)-dependent induction of a synthetic promoter

Near HSP70 Near infrared light stimulation Immunother Andersson infrared Heat promoter, of golden nanorods to activate apy of et al. light golden nanorods heat shock-responsive promoter cancer 201489

His-tagged TRPV1, Radio-wave heating of iron Radio Antibody-coated oxide nanoparticles causes Diabetes Stanley et Heat waves iron oxide calcium influx via TRPV1 treatment al. 201248 nanoparticles for leading to insulin expression His-tag

Spatially Red light inducible dimerization Red light Red light controlled TetR-PIF6 and and far red light-responsive Müller et and far red and far red engineering PhyB -VP16 dissociation to control gene al. 2013104 light light FR of expression angiogenesis

Cell encapsulation

The possibility of using a generic cell line, an allograft, or a xenograft rather than patient human leukocyte antigen (HLA)-compatible cells, could greatly increase the availability and reduce cost of cellular therapy. Unfortunately, each not fully compatible graft triggers an immune response, which eventually leads to its destruction. The immune rejection is primarily driven by the recognition of foreign antigens by T-cells, which subsequently destroy foreign cells. Traditionally, immune tolerance is imposed by persistent immunosuppressive therapy and associated with major side effects105. Cell encapsulation provides a physical barrier between transplanted cells and cytotoxic T-cells, and therefore, makes immune suppression redundant. Additionally, such a barrier prevents therapeutic cells from spreading across the organism and uncontrolled growth. The encapsulation method ensures long cell survival by enabling efficient diffusion of oxygen, nutrients, metabolites, as well as therapeutic protein. Inflammatory reactions caused by a foreign body could lead to production of potentially harmful mediators, as well as cause fibrotic tissue formation around the implant. Fibrosis limits the delivery of oxygen and nutrients to therapeutic cells, decreases their viability, and can potentially lead to immunomodulatory cytokines production106.

Page 20 of 200

Microencapsulation

Microencapsulation involves the immobilization of cells in polymeric capsules. Due to small capsule size, it offers very high surface-to-volume ratio, which facilitates the transport of oxygen, nutrients and metabolites. However, that makes them difficult to remove, which might pose a major risk in the case of malfunction. Historically, various concerns were raised that variable pore size and batch-to-batch differences in microencapsulation technologies do not ensure a sharp cutoff in molecule size, which might cause the destruction of therapeutic cells by circulating antibodies107 However, long term survival of microencapsulated cells suggests that humoral immune response is not a major concern108. Over the years, many natural and synthetic materials were applied for capsule formation109. Alginate, a polysaccharide compound from brown algae cell walls, is the most commonly used material due to its low cost, simple handling and early discovery. Bead formation is triggered by addition of multivalent cations, which form bridges between alginate molecules. The most frequently described protocol involves alginate core surrounded by a polycation layer (in. ex. poly-L-lysine, PLL) and an outer alginate layer. PLL increases the bead’s stability by creating strong interchain bridges and provides a biocompatible surface to increase cell survival109. Synthetic polymers may offer advantages comparing to natural compounds, such as greater mechanical resistance, chemical stability, higher batch-to-batch reproducibility, lack of immunostimulatory impurities and tunable properties110. Historically, they could not be used for mammalian cells encapsulation due to have harsh polymerization conditions, however multiple research groups developed materials and protocols to overcome that limitation110. Clinical trials with microencapsulated cells were performed mainly for type I diabetes treatment. Islets allografts or xenografts were efficiently protected from immune destruction without immunosuppression. The main limitation was long-term cell survival due to inflammatory response against the capsules and fibrotic tissue formation. Additionally, safety concerns were raised due to frequently occurring cell protrusion111. A schematic representation of microencapsulation is illustrated in Figure 3a.

Macroencapsulation

The production of macroencapsulation devices is separated in time from cell inoculation, which offers more flexibility in manufacturing process. Microfabricated membranes ensure highly uniform pore size distribution107. Additionally, solid implants have a higher mechanical durability, which decreases probability of cell protrusion to ensure safety. They can be placed in a defined spot of a body and easily removed. However, lower surface to volume ratio, comparing to microcapsules, and a lack of direct access to blood vessels limits

Page 21 of 200 the oxygen and nutrients diffusion and might limit cell viability. First experiments with macroencapsulation in Millipore membrane were performed in late 1940s by Algire et al. to study immune rejection108,112,113. Interestingly, already in these first studies, authors observed that transplant viability was compromised by fibrotic tissue formation. Consequently, a major focus in macroencapsulation technology development was put on material biocompatibility and its capability to induce neovascularization of the membrane-implant interface. Brauker et al. (1995) compared different membrane materials regarding the reaction of the host organism. They observed that bigger pore size (5μm) was associated with 80-100 fold more vascular structures than the smallest (0.02μm), even if the bigger-pore membrane was separated from cells with the small-pore membrane. Pore size which induced neovascularization was correlated with cell penetration114. Hence, double layer structure was used in later devices designed for automatic large scale fabrication115.

The problem of oxygen and nutrient availability was addressed by increasing the surface-to-volume ratio by designing flat sheet devices115,116 (Fig. 3b,c) , hollow fibers117-119, or wide-bore tubes108. Such designs were tested over years in animal models116,118-121 and even patients117. However, up to now they are considered insufficient for therapeutic applications, which even results in building prototypes with an external oxygen supply122. An alternative - intravascular implants or perfusion chambers increase the oxygen and nutrient supply by connecting therapeutic implants directly to blood vessels. Additionally, this solution facilitates efficient and long distance therapeutic protein transport and allows a faster reaction to changes in environment. Intravenously placed pancreatic islet-containing hollow fibers were tested in several type I diabetes animal models123-126 as well as human trials127, leading to restored normoglycemia. Unfortunately, implantation requires a complicated surgical procedure and is associated with a high risk of clot formation, which implies continuous anticoagulant therapy. Overall, it is frequently considered that the risks of intravascular implants outweigh the benefits128-130.

Page 22 of 200

Figure 3. Therapeutic cell encapsulation. (a) Microencapsulation. Therapeutic cells surrounded by a layer of alginate are protected from the host’s immune cells. The pore size allows for access of oxygen and nutrients, as well as secretion of metabolites and the therapeutic agent. (b, c) Macroencapsulation. Schematic representation of a flat sheet implant developed by Lathuiliere et al. 2015 115. Cells inhabit a hydrogel-filled cavity between two porous membranes. A supporting frame and reinforcement mesh provide mechanical durability. (b) A top view on the implant. (c) A side view of the implant. The scale bars represent the approximate size of the alginate bead or the implant. The scale between the bead/implant and other components presented in the panels is not maintained.

Page 23 of 200

Cellular therapies for type 1 diabetes

Diabetes mellitus type 1 (or type 1 diabetes, T1D) is an autoimmune disease leading to complete destruction of pancreatic β-cells resulting in persistent hyperglycemia131. Traditional treatment relies on regular subcutaneous injections of insulin or its derivatives with modified action time. Long-term remission can be achieved by islets transplantation132. Currently the major limitation is donor availability, since -cells are taken from healthy human subjects post mortem133. Therefore, developing glucose-responsive insulin-secreting cells would greatly increase treatment availability. Multiple stable β-cell lines have been developed, but none has been applied clinically134. Extensive efforts were especially directed towards the establishment of a protocol to recreate pancreatic islets efficiently by lineage control of autologous stem cells12,13. Potentially, such cells could be further modified to remove the target protein for autoimmune response. Such modification would be the key for curing diabetes, as otherwise transplanted autologous would cells be attacked by the patient’s immune system. An approach inspired by synthetic biology has been applied to create a β-mimetic cell line - HEK-β46. Glucose sensing in HEK-β was based on ATP-induced cell depolarization coupled to NFAT- driven transcription of modified insulin.

Islet encapsulation was recognized as an attractive method to avoid immune rejection of transplanted cells without administration of immunosuppressive drugs. Hence, that approach was extensively studied over the years128-130, leading to human tests in early 1990s117,135 and clinical trials106,136. In middle 2000s two companies announced clinical trials of encapsulated islet T1D in patients – Amcyte, using alginate microcapsules, and Novocell, with a flat sheet device136. In 2012-2014 Living Cell Technologies conducted phase 2b of clinical trials for their product Diabecell, but results were not published. In 2018 ViaCyte, a successor of Novocell continued phase I clinical trials for their products using PEC-01.

Page 24 of 200

Contribution of this work

This doctoral thesis focuses on the regulation of the behavior of therapeutic cells. The work described here has been presented in one published article and one original research manuscript. Chapter I introduces the first direct interface between actuating therapeutic cells and an electronic device. This interface helps bypassing a transcriptional delay by stimulating secretion of pre-produced insulin, which has been demonstrated in a clinically relevant model. Chapter II presents a toolbox of novel molecular regulators, which enable rewiring natural signaling pathways to alternative genomic targets for therapeutic cell reprogramming.

Contribution of CHAPTER I: “Electrogenetic cellular insulin release for real- time glycemic control in type-1 diabetic mice.”

In 1965, Gordon E. Moore projected that the number of components per integrated circuit would double every year. Moore’s “law” was proved remarkably prescient, with the rapid development of sophisticated and compact medical technologies, as a confirmation. Today, electronic wearables can monitor heart rate and glucose concentration, as well as gather and integrate a broad spectrum of diverse biometric data. Simultaneously, synthetic biology- inspired designer cells that respond to selected inputs with a therapeutic output were successfully implemented for treatment of a variety of diseases in rodent models. Therefore, a key challenge for next-generation synthetic biology-inspired treatment strategies is to combine these technologies by interfacing electronic devices with living designer cells in order to develop flexible, highly controllable therapeutic systems. Although several physical signals have been applied to control the behavior of therapeutic-producing cells, these systems rely on intermediates such as light or heat and have a high power requirement. However, the high power-efficiency associated with direct electrical stimulation would make this option more attractive.

The most impactful contribution of the work presented here lies in the creation of the first direct interface between therapeutic designer cells and an electronic device. Furthermore, this interface was used to develop a custom-built bioelectronic device that enables a very rapid electrogenetic control of insulin release. Specifically, (i) a direct, optogenetics-free, electrically inducible transgene expression system, (ii) a direct, optogenetics-free, electrically inducible fast insulin secretion system and (iii) a wirelessly controlled bio-electronic implant capable of rapid insulin secretion. This interface was built by genetically modifying cells to express both an L-type voltage-gated calcium channel and an inwardly rectifying potassium channel in order to Page 25 of 200 enable them to respond to electrical pulse stimulation. Electrostimulation depolarizes the cell membrane, opens the voltage-gated calcium channels, and triggers a calcium influx-mediated NFAT-dependent output. This system can be used to mediate either synthetic promoter-driven transgene expression or rapid vesicular secretion of constitutively expressed insulin.

Incorporating this system into pancreatic beta cell line, resulted in obtaining a cell line (Electroβ) that rapidly secretes insulin from vesicular stores in response to electrical stimulation. Peak secretion is attained within just 10 minutes. Furthermore, a subcutaneously implanted, custom- built bioelectronic device containing Electroβ cells could quickly restore normoglycemia in a type-1-diabetic mouse model. A key advantage of this approach is that it overcomes the problem of transcriptional delay associated with current synthetic biology-based type-1 diabetes treatment strategies.

Contribution of CHAPTER II: “Rewiring of endogenous signalling pathways to genomic targets for therapeutic cell reprogramming.”

Engineered cell-based treatment strategies hold a great promise to become one of the foundations of personalized medicine in the near future. The first clinically approved CAR-T cell therapies are already giving new hope to patients suffering from previously incurable leukemia and lymphoma. Numerous clinical trials and proof-of-concept research projects are underway to investigate the potential of engineered cells for treating cancer or chronic diseases or for organ regeneration. However, controlling the behavior of the therapeutic cells is critical and further progress requires precise and efficient methods to achieve this.

This chapter describes a novel approach for this purpose. The major advance provided by this work is a method of coupling extracellular signals to the transcription of alternative genes, thus avoiding the need to utilize artificial receptors. Specifically, it presents Generalized Engineered Activation Regulators (GEARs) – modular transcription activators that enable dCas9-driven rewiring of intracellular signaling pathways. The activity of GEARs is controlled by natural feedback loops, which enable gentle and fine-tunable responses. GEARs were used to redirect several signaling pathways important for pancreatic β-cell and immune cell functions. In particular, this system enables efficient membrane depolarization-induced activation of insulin expression in β-mimetic cells and immunostimulatory IL-12 expression in activated immortalized T cells (Jurkat), as well as simultaneous activation of multiple genes.

Page 26 of 200

CHAPTER I: Electrogenetic cellular insulin release for real-time glycemic control in type-1 diabetic mice

Krzysztof Krawczyk1,2, Shuai Xue1,3, Peter Buchmann1, Ghislaine Charpin-El-Hamri4, Pratik Saxena1, Marie-Didiée Hussherr1, Jiawei Shao3,5, Haifeng Ye3, Mingqi Xie1,5 and Martin Fussenegger1,6,*

1Department of Biosystems Science and Engineering, ETH Zurich, Mattenstrasse 26, CH-4058 Basel, Switzerland.

2Present Address: Novartis Pharma AG, P.O. Box, CH-4002 Basel, Switzerland.

3Shanghai Key Laboratory of Regulatory Biology, Institute of Biomedical Sciences and School of Life Sciences, East China Normal University, Dongchuan Road 500, Shanghai 200241, People’s Republic of China

4Département Génie Biologique, Institut Universitaire de Technologie Lyon 1, 74 Boulevard Niels Bohr, F-69622 Villeurbanne Cedex, France.

5Key Laboratory of Growth Regulation and Transformation Research of Zheijang Province, School of Life Sciences, Westlake University, Shilongshan Road 18, Hangzhou, People’s Republic of China.

6Faculty of Science, University of Basel, Mattenstrasse 26, CH-4058 Basel, Switzerland

*Corresponding author. E-mail: [email protected]

Science (in press)

Page 27 of 200

Sophisticated devices for remote-controlled medical interventions require an electrogenetic interface that uses digital electronic input to directly program cellular behavior. Here, we present a cofactor-free bioelectronic interface that directly links wireless-powered electrical stimulation of human cells to either synthetic-promoter- driven transgene expression or rapid secretion of constitutively expressed protein therapeutics from vesicular stores. Electrogenetic control was achieved by coupling ectopic expression of L-type voltage-gated channel (CaV1.2) and inwardly rectifying potassium channel (Kir2.1) to the desired output through endogenous calcium signaling.

Focusing on type-1 diabetes, we engineered electro-sensitive human β-cells (Electroβ).

Wireless electrical stimulation of Electroβ inside a custom-built bioelectronic device provided real-time control of vesicular insulin release; insulin levels peaked within 10 minutes. When subcutaneously implanted, this electro-triggered vesicular release system restored normoglycemia in type-1-diabetic mice.

Electronic insulin expression and electro-stimulated insulin release for next-generation diabetes therapy.

Page 28 of 200

Precise control of dosage is essential for the success of any drug-based therapy 137-140. However, taking pills or administering biopharmaceuticals at regular intervals based on body weight, as is standard medical practice, is far from being precise and does not reflect the dynamics required for sophisticated metabolic interventions 137-140. Cell-based therapies capitalizing on implanted encapsulated designer cells engineered to fine-tune in-situ production and systemic delivery of protein therapeutics in response to chemical and physical cues have shown promising results in proof-of-concept studies 16,46. Since chemical control input is often limited, traceless physical cues such as light (optogenetics) 17,47,57,93, heat transmitted by magnetic fields (magnetogenetics) or radio waves (radiogenetics) 49,89-91 are attractive for achieving rapid remote-control of therapeutic transgene expression, because they avoid the side effects 141,142 and challenges with bioavailability or pharmacodynamics of chemical trigger compounds 143-146. However, available physically triggered gene switches may require a high energy input 16,47,57, often involve complex chemical or inorganic cofactors 48,90, and may require fine-tuning of the transcription of the therapeutic transgenes, which slows down the overall response dynamics 16,46,57,90,147. Thus, direct cofactor-free wireless electrical stimulation of engineered cells to control vesicular secretion of protein therapeutics in a robust, adjustable and repeatable manner would offer substantial advantages for medical applications by enabling direct communication between electronic devices and designer cells.

Although cellular metabolism and human-made electronics share similar operating principles in terms of input sensing, information processing and output production, the core information transfer and processing functions of living and electronic systems are different, which limits their interoperability. Humans employ ion gradients across insulated membranes to simultaneously process slow analog chemical reactions and communicate information in multicellular systems through soluble or gaseous molecular signals. In contrast, electronic systems use multicore central processing units to control the flow of electrons through insulated metal wires with gigahertz frequency and communicate information across networks via wired or wireless connections. Thus, direct electrical stimulation of gene expression or vesicular secretion requires a bioelectronic interface that manages electrical conduction between electrodes and electrosensitive designer cells, as well as conversion of electronic information via depolarization to protein production and release.

The first attempts to create an electrogenetic interface were reported over a decade ago 95,148, but that interface was neither direct nor usable under physiological conditions. More recently, a SoxR-based redox system that can control gene expression in Escherichia coli was

Page 29 of 200 reported 98, but this was also not direct, and was too toxic for in-vivo application. Thus, despite decades of expertise in converting trigger-inducible bacterial and fungal repressor-operator interactions into synthetic mammalian gene switches, simple translation of bacterial electrogenetics into a mammalian cellular context has been unsuccessful due to the cytotoxicity, limited bioavailability and poor clinical compatibility of electro-sensitive redox compounds 95.

With the advent of optogenetics it became possible to remote-control target gene expression by illumination with light and so to indirectly link electrical stimulation via a light source with cellular transcription control 16,17. This enabled glycemic control of experimental type-2 diabetes by controlling an optogenetic biomedical implant with a smartphone to upload instructions for designer cells to produce and systemically deliver a therapeutic dose of an insulinogenic peptide 16. However, the optogenetic device requires a considerable amount of energy to operate the light source 16,17. The power-efficiency associated with direct electrical stimulation is a major reason why clinically licensed pacemakers can be battery-powered for a lifespan of at least 15 years 149. Other major challenges to the clinical application of optogenetic technology include illumination-based cytotoxicity 150, the use of bacterial components 16,17,144- 146, and the need for sophisticated chemical or inorganic cofactors that have side effects 151-153, poor bioavailability or short half-lives in vivo 154. Other traceless physical control technologies based on electro-induced heat transmission such as magneto- and radiogenetics share the same challenges 48,90,155,156.

Diabetes is a common, chronic condition, and so is an attractive target for individualized precision treatment. Regulation of blood-glucose levels is a closed-loop homeostatic process. Glucose-stimulated insulin release by pancreatic β-cells involves uptake and metabolism of glucose, ATP-mediated closure of potassium channels, depolarization of the plasma membrane, and opening of the voltage-gated calcium channels, which results in an intracellular Ca2+ surge and concurrent rapid release of insulin from intracellular storage vesicles 157. For intervention in this process, we aimed to design a bioelectronic interface consisting of an implantable platform that combines electronics and electrosensitive designer cells that can release insulin on demand. The implant would incorporate a cell chamber containing semipermeable membranes that permit nutrient supply and product delivery via fibrous connective tissue, while protecting the designer cells from cellular host responses 158,159 and securely containing them for safety reasons 160. To address this need, we describe here a direct cofactor-free electrogenetic interface to trigger vesicular secretion of insulin by using electrical stimulation to modulate the membrane polarization of human -cells (Electro) engineered for ectopic

Page 30 of 200 expression of calcium and potassium channels. Furthermore, to validate our approach, we incorporated these electro-sensitive designer cells into a bioelectronics implant and evaluated its performance in a mouse model of type-1 diabetes.

Results

Membrane depolarization-based transcriptional control in mammalian cells

L-Type voltage-gated calcium channels consist of α1, α2, δ, and β subunits and are essential for the functioning of cardiomyocytes, neurons and endocrine cells 161. These channels open upon membrane depolarization, and the resulting calcium influx regulates muscle contraction, vesicular secretion of hormones, and NFAT (nuclear factor of activated T-cells)- driven induction of target genes 162.

To design a mammalian transcription-control circuit responsive to membrane depolarization, we cotransfected HEK-293T cells with one of the three L-type voltage-gated calcium channels, CaV1.2, CaV1.342A or CaV1.342, encoded by the common α2/δ1 (pCaVα2δ1,

PhCMV-α2/δ1-pA) and β3 (pCaVβ3, PhCMV-β3-pA) subunits and the respective channel-forming subunit α1C (pCaV1.2, PhCMV-α1C-pA), α1D42A (pCaV1.342A, PhCMV-α1D42A-pA) or α1D42

(pCaV1.342, PhCMV-α1D42-pA), as well as the reporter plasmid pMX57 encoding the human placental secreted alkaline phosphatase (SEAP) driven by the PNFAT3 promoter (pMX57,

PNFAT3-SEAP-pA) (Fig. 1A). Depolarization of channel-transgenic HEK-293T cells with 40 mM KCl revealed that ectopic expression of CaV1.2 showed the highest depolarization- triggered SEAP induction (Fig. 1B).

Coexpression of the inwardly rectifying potassium channel Kir2.1 (pKir2.1, PhCMV-

Kir2.1-pA), which has been reported to decrease the resting membrane potential of mammalian cells 163, substantially decreased basal SEAP expression and improved the overall induction profile of the depolarization-triggered CaV1.2-mediated transcription-control device (Fig. 1C).

Combinatorial analysis of the importance of CaV1.2’s individual components for overall depolarization-triggered transcription control revealed that the channel-forming α1C subunit was essential, whereas α2/δ1 and β3 were not, although their absence reduced the maximum

SEAP expression (Fig. S1). Therefore, we used cells expressing the full CaV1.2 with the α1C,

α2/δ1 and β3 components as well as Kir2.1, referred to as ElectroHEK, in all follow-up experiments. Importantly, ElectroHEK cells are not activated by physiological ion concentrations, not even at life-threatening levels of KCl (6.5 mM (27)) or at CaCl2 levels representing a medical emergency (3.5 mM (28)) (Fig. S2). Page 31 of 200

Page 32 of 200

Figure 1. Design of the electrogenetic circuit in mammalian cells. (a) Schematic representation of the electrogenetic circuit. The inwardly rectifying potassium channel lowers the resting membrane potential of HEK293T cells and electrical pulses depolarize the plasma membrane and open the L-type voltage-gated calcium channel. Calcium influx activates the calmodulin/calcineurin pathway, which leads to dephosphorylation of NFAT and its translocation to the nucleus, where it activates the NFAT-sensitive promoter and triggers transgene expression (b) Comparative performance of three L-type voltage-gated calcium channels. Cells were co-transfected with PNFAT3-driven SEAP reporter plasmid (pMX57), plasmids encoding, α2/δ1 (pCaVα2δ1, PhCMV-α2/δ1-pA) and, β3 (pCaVβ3, PhCMV-β3-pA), and one of the pore-forming subunits: α1C (pCaV1.2, PhCMV-α1C-pA), α1D42A (pCaV1.342A, PhCMV-

α1D42A-pA) and α1D42 (pCaV1.342, PhCMV-α1D42-pA), to form CaV1.2, CaV1.342A, and

CaV1.342 accordingly. pcDNA3.1(+) was used as a mock plasmid. The cell membrane was depolarized with 40 mM potassium chloride (red bars) and after 24 hours SEAP was quantified in the supernatant. Blue bars show negative controls. (c) Co-expression of L-type voltage-gated calcium channel CaV1.2 and inwardly rectifying potassium channel Kir2.1. Cells were co- transfected with pCaV1.2 (PhCMV-α1C-pA), pCaVα2δ1 (PhCMV-α2/δ1-pA), pCaVβ3 (PhCMV-β3-pA), pKK05 (PhCMV-Kir2.1-pA) and pMX57 (PNFAT3-SEAP-pA) in the molar proportions 1:1:1:1:3. Cells were depolarized with 40 mM KCl for 24 hours (red bars) and SEAP was quantified in supernatant samples. Bars represent mean ± SEM. n = 3. **p<0.01, ***p<0.001.

Design and characterization of a synthetic electrogenetic mammalian transcription-control device

To test whether transgene expression can be directly triggered by electrically stimulated membrane depolarization, we electrostimulated the ElectroHEK cells transfected with the PNFAT3- driven SEAP expression vector (pMX57, PNFAT3-SEAP-pA), using voltage-controlled square unipolar pulses with alternate polarization 164-166 (Fig. 2A). Indeed, electric pulse stimulation triggered pMX57-transgenic ElectroHEK cells to produce high levels of SEAP (Fig. 2B-D). The electrostimulated transgene expression could be fine-tuned by voltage (maximum SEAP induction at 50 V) (Fig. 2B) as well as adjusted by altering the pulse length (maximum SEAP induction at 2 ms) (Fig. 2C). Full activation of the system was reached after 4 hours of stimulation (Fig. 2D). Electrostimulation efficiency did not depend on the pulsing frequency within the range of 0.5-10 Hz (Fig. 2E). Importantly, the parameter set employed for effective electrostimulation did not decrease cell viability (Fig. S3A-D). Additionally, CaV1.2-deficient HEK-293T cells were insensitive to electrostimulation (Fig. S3E). Kinetic experiments revealed maximum SEAP expression 7 hours after the beginning of stimulation (Fig. S4A) and confirmed the reversibility of the system (Fig. S4B).

Page 33 of 200

Page 34 of 200

Figure 2. Characterization of the electrogenetic circuit in vitro. (a) Schematic representation of electrical stimulation setup. Cells were stimulated with carbon hanging electrodes (C-dish) using monopolar pulses with alternate polarization. (b-e) Cells were co-transfected with pCaV1.2 (PhCMV-α1C-pA), pCaVα2δ1 (PhCMV-α2/δ1-pA), pCaVβ3 (PhCMV-β3-pA), pKK05 (PhCMV-

Kir2.1-pA and pMX57 (PNFAT3-SEAP-pA). SEAP assay was performed 24 hours after the beginning of the electrical stimulation procedure. Blue bars represent unstimulated control, orange bars show electrically stimulated samples, and red bars indicate cells depolarized with 40 mM KCl. (b) Voltage dependence. Electrical stimulation was performed for 1 hour with 2 ms pulses at 10 Hz frequency and the indicated voltage. (c) Pulse length effect. Electrical stimulation was performed for 1 hour at 50 V, 10 Hz frequency, and the indicated pulse length. (d) Time course. Electrical stimulation was performed for the indicated period of time with 2 ms pulses at 0.5 Hz and 50 V. (e) Frequency effect. Electrical stimulation was performed for 1 hour with 2 ms pulses at 50V and at the indicated frequency. Bars represent mean ± SEM. n = 3. Statistical significance was calculated versus the negative control. ns – not significant, *p<0.05, **p<0.01, ***p<0.001.

Design of the bioelectronic implant

Translation of electrostimulated gene expression into a clinical proof-of-concept bioelectronic implant required a more compact design for electrodes and electrostimulation. Simple miniaturization of the free-hanging electrodes used in the device described above did not provide efficient electrostimulation. Thus we designed a custom-engineered cell-culture insert containing electrodes on either side of a semipermeable membrane harboring a monolayer of electrosensitive ElectroHEK cells (Fig. 3A). Electrostimulation of pMX57 (PNFAT3-SEAP-pA)- transfected ElectroHEKs resulted in peak SEAP levels at 7.5 V (Fig. 3B,C), which is one order of magnitude lower than that of the previous free-hanging electrode arrangement, and at shorter pulse length (Fig. 3D,E); both factors are important for high power efficiency of any electrostimulation device.

Page 35 of 200

Page 36 of 200

Figure 3. Design and functionality of the bioelectronic implant in vitro (a) Schematic representation of the stimulation setup in a cell culture insert. Two platinum electrodes (blue) were placed on opposite sides of the porous membrane covered with cells, and electrical pulse stimulation was applied. SEAP was quantified 24 hours after stimulation in the supernatants of the cell-culture insert (above the membrane) and the well of the cell-culture plate (below the membrane), to confirm that the secreted protein diffuses across the membrane of the cell-culture insert. (b, c) Voltage-dependent response of electrically stimulated pMX57-transfected ElectroHEK cells grown in a cell culture insert. Cells were stimulated with 2 ms pulses at 1 Hz frequency for 1 h (orange bars). SEAP was measured in supernatant samples from the cell- culture insert (above the membrane) (b) and from the cell-culture well (below the membrane) (c). Blue bars show the negative control. (d, e) Pulse length-dependence. Cells were stimulated with 7.5 V pulses at 1 Hz frequency for 1 h (orange bars). SEAP was measured from supernatant samples from above (d) and below the cell layer (e). Blue bars show the negative control. Bars represent mean ± SEM. n = 3. Statistical significance was calculated versus negative control. ns – not significant, *p<0.05, **p<0.01, ***p<0.001.

To enable electrostimulated transgene expression by electrosensitive cells in vivo, we designed a wireless-powered bioelectronic implant. The custom-engineered cell-culture insert equipped with the electrodes was clicked into a 3D-printed FDA-licensed polyamide casing (Fig. 4a, c) containing a sealed electronic switchboard (Fig. S5, S6) that generates the square unipolar pulses for electrostimulation of the encapsulated ElectroHEKs. The implant’s electronics is inductively powered and controlled by an extracorporeal field generator that wirelessly communicates with the bioelectronic implant at the ISM (industrial, scientific and medical) frequency (13.56 MHz) (Fig. 4B, S7, S8). The implant generates square pulses, the voltage of which is dependent on the distance to the center of the field generator (Fig. S9). The electronic circuit is insensitive to temperatures between 25 C and 50 C (Table S1). A control run of the bioelectronic implant validated wireless-controlled electrostimulated SEAP expression of pMX57-transfected ElectroHEK cells (Fig. 4C). We confirmed that the bioelectronic implants are IPX7 waterproof (International Protection Marking, IEC standard 60529) and show no cell leakage in a five-day in vitro experiment (Table S2).

Page 37 of 200

Page 38 of 200

Figure 4. Bioelectronic implant in vitro (a) Three-dimensional model of a disassembled bioelectronic implant. A ring containing a porous membrane on one side can be assembled with a 3D-printed polyamide frame to form a cell chamber. The electronic switchboard is placed on the other side of the frame. The active platinum electrode (placed in the cell chamber; invisible in the model) is soldered to a connector on a switchboard. The ground electrode, made out of thin stainless steel mesh, is connected to the second connector on a switchboard. (b) The bioelectronic implant can be placed subcutaneously on the dorsal side of the mouse, with the cell chamber facing down. The field generator provides wireless energy transmission. A red diode enables implant function monitoring. (c) Photograph of two bioelectronic implants with a coin for comparison. (d) Comparison of external generator-powered and implant-powered electrostimulation of pMX57-transfected ElectroHEK cells. SEAP was measured in supernatant samples from above the cell layer. n=3. Data points represent mean ± SEM. Statistical significance was calculated versus negative control. ***p<0.001.

Electroβ cells providing electrostimulated vesicular secretion

Since ElectroHEK-based insulin production is transcription-based, it lacks the rapid release dynamics of vesicular secretion characteristic of native pancreatic β-cells 46. To engineer mammalian cells for electrostimulated vesicular release of insulin (Fig. 5A), we derived a 167 monoclonal population INSVesc from the pancreatic β-cell line 1.1E7 by selection for deficiency in glucose sensitivity (Fig. 6E,F), but with retention of the vesicular insulin- secretion machinery. Indeed, electron micrographs of Electroβ, an INSVesc variant stably transgenic for constitutive expression of CaV1.2 and Kir2.1 channels (pKK66, PhEF1α-α1C-P2A-

Kir2.1-pA; pMX251, PhEF1α-α2/δ1-P2A-β3-pA) as well as Proinsulin-NanoLuc, a designer construct engineered to co-secrete insulin and the Oplophorus gracilirostris luciferase (NanoLuc) at an equimolar ratio in endocrine cell types 168 (Fig. 5A), revealed storage vesicles reminiscent of insulin-containing granules of human islet-derived β-cells (Fig. 5D,E).

Additionally, Electroβ showed well-correlated vesicular insulin and NanoLuc secretion in response to KCl-mediated (Fig. 5B,C) or electrostimulated (Fig. 6A,B) membrane depolarization. The stability and functionality of the Electroβ cell line were confirmed over at least 30 passages during 3 months in continuous culture (Fig. S10).

Page 39 of 200

Page 40 of 200

Figure 5. Electrogenetic engineering of β-cells. (a) Schematic representation of the electrically inducible insulin secretion pathway. The inwardly rectifying potassium channel

Kir2.1 lowers the resting membrane potential, which keeps the voltage gated calcium channel

CaV1.2 closed. Electrical pulse stimulation causes membrane depolarization, opening of CaV1.2 and calcium influx, which stimulates vesicle secretion. Vesicles are loaded with pre-produced insulin (red dots) and NanoLuc (yellow dots). (b, c) Comparison of insulin secretion by INSVesc and Electroβ cells. Vesicle secretion was quantified by insulin-specific ELISA (b) and luminescence (c) before (blue bars) and after depolarization with 40 mM KCl (red bars). BDL

– below detection limit. (d) Transmission electron microscopy (TEM) pictures of Electroβ cells. White arrows indicate insulin-containing vesicles. (e) TEM pictures of primary β-cells from human pancreatic islets. White arrows indicate insulin-containing vesicles. Bars represent mean ± SEM. n = 3. ***p<0.001.

The depolarization-based insulin-release dynamics was profiled by electrostimulating

Electroβ and recording the corresponding NanoLuc-mediated luminescence in the culture supernatant (Fig. 6C). Peak NanoLuc levels were reached within ten minutes following electrostimulation (Fig. 6C), compared to 8 hours for transcription-based insulin production 46 57 and secretion by ElectroHEK, HEK-β and OptoHEK (Fig. S11). When repeatedly electrostimulated, Electroβ recovered full secretory capacity after four hours (Fig. 6D). Most importantly, Electroβ did not show any glucose-sensitive insulin production, which ensures exclusive electrostimulation control of vesicular insulin secretion, without interference from blood-glucose levels (Fig. 6E,F). Overall, Electroβ showed similar electrostimulation parameters to ElectroHEK (Fig. S12A-D). To illustrate the broad applicability of our approach, we also demonstrated electro-stimulated vesicular secretion of glucagon by pancreatic alpha cells, which secrete the insulin counter-regulatory hormone glucagon by calcium-triggered vesicular release 169 (Fig. S13) – this is akin to the β-cell-mediated insulin secretion.

Page 41 of 200

Page 42 of 200

Figure 6. Functionality of Electroβ cells in vitro. (a, b) Electrostimulation of Electroβ cells. Cells were seeded into cell culture inserts, and 24 hours later they were stimulated with electrical pulses (orange bars) or with 40 mM KCl (red bars). Blue bars show the negative control. n = 3. (a) Insulin content in the supernatant from inside the insert (above the cell layer) was measured by ELISA. (b) Luminescence measured in supernatant samples taken from inside the insert (above the cell layer). n = 3. (c) Secretion kinetics. Electroβ cells were seeded into cell culture inserts and stimulated with electrical pulses (red frame). Luminescence was measured in supernatant samples every 10 min. n = 4 (d) Reversibility assay. Electroβ cells were electrostimulated for 15 min twice, with 4 h time intervals between the first and second electrostimulation. n = 4 (e) Glucose-induced insulin release. Electroβ cells were incubated with various concentration of glucose for 15 minutes (blue bar – 2.8 mM glucose; orange bars – elevated glucose; red bar – 2.8 mM glucose with 40 mM KCl). Luminescence was measured in supernatant samples. n = 3. (f) Glucose-induced insulin release. INSVesc cells were incubated with various concentration of glucose for 60 minutes (blue bar – 2.8 mM glucose; orange bars – elevated glucose; red bar – 2.8 mM glucose with 40 mM KCl). Insulin content was quantified in supernatant samples. n = 3. Bars and dots represent mean ± SEM. Statistical significance was calculated versus negative control. ns – not significant, *p<0.05, **p<0.01, ***p<0.001.

Wireless electrostimulated vesicular secretion of insulin provides rapid glycemic control in type-1 diabetic mice

Native pancreatic β-cells release the insulin stored in granules via a process known as vesicular secretion 157. The immediate release of stored insulin improves the response dynamics and rapidly restores blood-glucose homeostasis in response to postprandial excursions. So far, designer cell-based proof-of-concept strategies to treat experimental diabetes have focused on transcriptional control, which is considered too slow to cope with postprandial blood-glucose surges 16,46,48,49,90. For example, previously reported HEK-β cells 46, which rely on transcriptional control and the classical secretory pathway for insulin release, require up to 24 hours to reach physiological blood-insulin levels (Fig. S14). Similar performance was observed 57 for OptoHEK . In contrast, when placed into the wireless-powered bioelectronic implant (Fig.

4), Electroβ cells could re-establish postprandial glucose metabolism in insulin-deficient type-1 diabetic mice following a brief electrostimulation without causing hypoglycemic excursions (Fig. 7A), and instantaneously decrease blood-glucose levels to restore normoglycemia following electrostimulation (Fig. 7B). Notably, the results of glucose tolerance tests revealed comparable performance between Electroβ and human pancreatic islets, which are known to release insulin by vesicular secretion upon glucose sensing (Fig. S7A). Fast vesicular secretion was also confirmed by blood-luminescence quantification 170, which showed a peak signal just Page 43 of 200

1 hour following electrostimulation, returning to baseline after 2 hours (Fig. 7C). Glycemia could also be controlled over longer periods of time without any sign of hypoglycemia (Fig. 7D).

Page 44 of 200

Figure 7. Comparative analyses of Electroβ-containing bioelectronic implants in type-I diabetic mice. Type-1-diabetic mice implanted on the back with Electroβ-containing bioelectronic devices were profiled for blood-glucose dynamics. (a) Glucose tolerance test. At 48 hours after implantation, the Electroβ cells inside the bioelectronic implant were electro- stimulated for 60 min (red line), then the animals were given intraperitoneal glucose injections and their blood-glucose levels were monitored. All groups received intraperitoneal glucose injection (2 g per kilogram of body weight). Wild type (n = 8); T1D, implant electrostimulated (type I diabetes, activated implant, n = 6); T1D, empty implant (type-I diabetes, implant without cells, n=10), islets (human pancreatic beta islets, n = 3). The statistical significance of differences between the electrostimulated and the mock group was calculated. (b) Real-time glycemia measurement. Fasted type-1-diabetic mice implanted with Electroβ-containing bioelectronic implants were electro-stimulated for 30 minutes and their glycemic profile was recorded. n = 6 for the non-stimulated control (T1D, implanted mice), n = 7 for the stimulated group (T1D, implanted mice) and n = 6 for wild-type control. The green frame indicates the normoglycemic range (4.4 - 7.2 mM). The statistical significance was calculated between the electrostimulated and the non-stimulated control. (c) Blood-luciferase kinetics of animals implanted with Electroβ-cell-containing implants electro-stimulated for 30 min (red line; n = 6). NanoLuc was quantified from microliter-scale blood samples every 30 minutes. The blue line indicates the non-electro-stimulated negative control (n = 5). The statistical significance of differences versus time point 0 was calculated with a paired t-test. (d) Fasting glycemia. Type- 1-diabetic mice were implanted with Electroβ-containing bioelectronic implants and fasting glycemia was recorded for over a week. Orange line indicates the initial level of average glycemia. The statistical significance of differences versus time point 0 was calculated with a paired t-test. Red frames indicate electrostimulation time. ns – not significant, *p<0.05, **p<0.01, ***p<0.001.

Biocompatibility and functional longevity of the bioelectronic implant

To validate the biocompatibility of the bioelectronic implants, we analyzed treated animals as well as explanted devices at three weeks after implantation, according to ISO 10993 171, and we observed no material cytotoxicity, systemic kidney or liver toxicity, or alteration of hematologic profile or systemic immune responses; in addition, we saw no local immune-cell infiltration or substantial fibrotic tissue formation at the implant-tissue interface. There was no apparent indication of implant-related cytotoxicity (Fig. S15) or systemic toxicity (Table S3), and no apparent difference in hematologic profiles among cell-containing and cell-free bioelectronic implants and biocompatible control implants (Table S4). Likewise, we found no marked difference in the well-vascularized fibrous capsule surrounding the implants (Fig. S16) or in immune-cell infiltration (Fig. S17, Table S5) among cell-containing, cell-free and

Page 45 of 200 biocompatible control implants. Mice implanted with Electroβ-cell-containing bioelectronic devices showed no change of body weight compared to untreated animals; also, signs of irritation or inflammation, as well as serum levels of inflammatory cytokines, were similar to or lower than those of animals treated with cell-free or biocompatible reference implants (Fig. S18, Fig. S19). Visual inspection of explanted bioelectronic devices showed no decomposition and no apparent erosion (Fig. S20).

In view of the need for clinical translation towards a lifestyle-compatible therapeutic product, we adapted the bioelectronic implant architecture to allow repetitive exchange of individual cell batches over time (Fig. S20A). Sequential in-situ “refilling” of the implanted bioelectronic device with fresh batches of Electroβ-cells without the need for surgical removal or replacement of the implant will reduce cost as well as implant-associated infections, while increasing patients’ convenience and treatment longevity. Insulin levels of type-1-diabetic mice, which had the Electroβ cells of their bioelectronic implants replaced once a week for a period of three weeks, were restored after remote-controlled electro-stimulated insulin release by Electroβ cells (Fig. S20B, C). Together, these results suggest that the bioelectronic implant successfully integrates the advantages of electronics-based 172 and cell-based counterparts 46, and represents a promising approach to diabetes treatment.

Discussion

In this work, we have eliminated the need to use light as a converter between electronics and genetics, advancing optogenetics into electrogenetics by engineering a direct, co-factor- free electrogenetic interface that enables electronics to directly program gene expression as well as vesicular secretion in human cells. Furthermore, by incorporating electrogenetic designer cells (Electro) containing this interface into a bioelectronic implant, we have successfully implemented a proof-of-concept device providing rapid electro-stimulated insulin release for the treatment of experimental type-1 diabetes. The overall slow response dynamics associated with transcription-based control systems 16,17,46,48,57,90,141,142,144-146 highlights the importance of vesicular secretion for the treatment of diabetes, which requires quick vesicular release of insulin to respond rapidly to postprandial blood-glucose surges 173,174. Indeed, we found that wireless electrical stimulation of vesicular insulin release from our engineered Electroβ cells encapsulated in a bioelectronic implant could attenuate postprandial hyperglycemia in type-1 diabetic mice with comparable performance to transplanted human pancreatic islets.

Page 46 of 200

Taking account of the importance of economical manufacturing, we integrated all components of the bioelectronic implant into a 3D-printed polyamide casing. Although the bioelectronic implant could in principle be powered by batteries 175 (Table S4), for practical reasons, including the limited space for implantation and the intrusiveness of animal experimentation, we chose to power the device inductively at 13.56 MHz, an FCC-licensed radio frequency that is reserved internationally for industrial, scientific and medical devices and does not interfere with telecommunications. Due to the power efficiency of the implant, we speculate that wireless-powered control by wearable devices such as smartphones and smartwatches might be feasible in the near future.

However, reaching the full therapeutic potential of electrogenetics will require closed- loop control. Whereas classical medical interventions are open-loop, since the dose is largely determined by the physician based upon body weight, closed-loop systems enable feed-back control that coordinates biomarker input to therapeutic output and provides an autonomous and self-sufficient interface with patients’ metabolism. For electrogenetic type-1 diabetes control, this would mean using electronic blood-glucose sensors to directly control electro-stimulated insulin release in real time, much like the concepts currently being explored for prototypes of the bionic pancreas 176. However, electronic closed-loop systems operating in the bionic pancreas require frequent calibration, and have a short life-span of only a few days 172. On the other hand, incorporation of a microcontroller and/or a glucometer into our bioelectronic implant to achieve closed-loop insulin control should be a straightforward electrical engineering implementation. Most importantly, the delayed resorption of insulin from subcutaneous tissues to which insulin is delivered by the bionic pancreas requires dual-hormone control using glucagon to counteract or prevent insulin-mediated hypoglycemia 177,178. We show here that glucagon can be released from pancreatic α-cells by vesicular secretion, just as insulin is from β-cells, suggesting that a dual-hormone electrogenetic system using two types of engineered cells would be feasible. Nevertheless, dual-hormone control is not expected to be necessary with our electrogenetic system, because, as noted above, the dynamics of electro-stimulated vesicular insulin secretion from Electroβ cells appear to be comparable with those of human pancreatic islets. Furthermore, the demonstration that our system works in two different types of cells suggests broad potential applicability of electrogenetics for electro-stimulated hormone release in future cell-based therapies

As in the case of the bionic pancreas 176, long-term functionality of cellular implants remains a major challenge in designing next-generation encapsulated cell-based therapeutic

Page 47 of 200 devices 136. A recent clinical trial using encapsulated pancreatic progenitor cells, the precursor phenotype of insulin-secreting β-cells (Viacyte’s VC-01), confirmed the need for further technological development to promote engraftment 179 Long-term functionality of cells inside implants remains among the challenges facing translation of academic proof-of-concept studies into clinical reality. In this context, the first initiatives to improve viability (Beta-O2 Technologies - ßAir) as well as vascularization of encapsulated cells 122 (e.g., Viacyte’s PEC Direct or Sernova’s Cell Pouch System) have already begun in industry.

We have shown here that wireless electrical stimulation of insulin release by electro- sensitive designer cells inside a bioelectronic implant was able to rapidly restore normoglycemia in type-1 diabetic mice. Wireless electronic devices programming the release of biopharmaceuticals, either via the secretory pathway or vesicular secretion, by means of direct communication between electronic devices and implanted cells is expected to open up many new opportunities for advanced precision healthcare optimized for individuals.

Page 48 of 200

Acknowledgments: We would like to thank the Geneva Islet Transplantation Center and Henryk Zulewski for human pancreatic islets, which were obtained through the basic research program of the European Consortium for Islet Transplantation (ECIT) supported by the Juvenile Diabetes Research foundation (JDRF; grant no 31-2008-416). We would like to thank Sebastian Bürgel for help in the initial stage of the project, Andreas Hierlemann for providing pulse generators, Marc Folcher for constructive discussions, Erdem Siringil for support with 3D printing, Brian Lang for advice on statistical analysis, Ana Margarida Palma Teixeira and Gieri Camenisch for preparation of the animal experimentation applications, Alexandra Graff-Meyer and Christel Genoud for taking electron microscopy images, and Haijie Zhao and Nik Franko for their help in assembling implants.

Funding: This work was supported by a European Research Council (ERC) advanced grant (ElectroGene; grant no. 785800) and in part by the Swiss National Science Foundation (SNF) National Centre of Competence in Research (NCCR) for Molecular Systems Engineering.

Author contributions: KK and MF designed the project. KK and PS performed the cell culture experiments and KK designed the implants. KK, SX, GC, MDH, JS and HY performed the animal experiments. PB designed the electronic switchboard. KK, PS, MX and MF designed the experiments and analyzed the results. KK, PS, MX and MF wrote the manuscript. SX, MDH, MX and MF designed the modified implants and the in vivo “refill” as well as the insulin kinetics experiments.

Competing interests: The authors declare no competing financial interests.

Data and materials availability: The authors declare that all the data supporting the findings of this study are available within the paper and its supplementary information file. Original plasmids are available upon request. All vector information is provided in Table S7.

Supplementary Materials:

Materials and Methods

Figures S1-S20

Tables S1-S8

Page 49 of 200

Supplementary Materials

Electrogenetic cellular insulin release for real-time glycemic control in type- 1 diabetic mice

Krzysztof Krawczyk, Shuai Xue, Peter Buchmann, Ghislaine Charpin-El-Hamri, Pratik Saxena, Marie-Didiée Hussherr, Jiawei Shao, Haifeng Ye, Mingqi Xie and Martin Fussenegger

Correspondence to: [email protected]

This PDF file includes:

Materials and Methods

Figs. S1 to S20

Tables S1 to S8

Page 50 of 200

Materials and Methods

Key plasmids used in this study. Comprehensive design and construction details for all expression vectors are provided in Table S7. Key plasmids include (i) pCaVα1C (PhCMV-α1C- pA), pCaVα2δ1 (PhCMV-α2/δ1-pA) and pCaVβ3 (PhCMV-β3-pA), which enable constitutive expression of subunits α1C, α2, δ1 and β3 of the L-type voltage-gated channel CaV1.2, 2+ respectively (ii) pMX57, which encodes a Ca -responsive PNFAT3-driven SEAP expression unit

(PNFAT3-SEAP-pA) (5), (iii) pKK66, which harbors a Sleeping Beauty (SB) 100X-specific transposon containing a bicistronic unit for PhEF1α-driven expression of the CaV1.2 α1C subunit and the inwardly rectifying potassium channel Kir2.1, as well as a bicistronic unit for the expression of the blue fluorescent protein (BFP) and the puromycin resistance gene (PuroR)

(ITR-PhEF1α-α1C-P2A-Kir2.1-pA-ITR:PRPBSA-BFP-P2A-PuroR-pA-ITR), (iv) pMX251, which harbors a SB100X-specific transposon containing a constitutive bicistronic unit for the expression of the CaV1.2 α2, δ1 and β3 subunits, as well as a bicistronic unit for constitutive expression of the red fluorescent protein dTomato and the blasticidin resistance gene (BlastR)

(ITR-PhEF1α-α2/δ1-P2A-β3-pA:PRPBSA-dTomato-P2A-BlastR-pA-ITR) (5), and (v) pProinsulin- NanoLuc (PhCMV-Proinsulin-NanoLuc-pA), a lentiviral expression vector containing a modified proinsulin whose C peptide has been replaced with Oplophorus gracilirostris luciferase (NanoLuc)168.

Cell culture and transfection. Human embryonic kidney cells (HEK-293T, ATCC: CRL-11268) were cultivated in Dulbecco’s modified Eagle’s medium (DMEM; cat. no. 52100- 039; Thermo Fischer Scientific, Waltham, MA, USA). 1.1E7 cells (cat. no. 10070101-1VL, Sigma-Aldrich, Saint Louis, MO, USA) were cultivated in Roswell Park Memorial Institute 1640 medium (RPMI; cat. no. 72400-021, Thermo Fischer Scientific) supplemented with 10% fetal bovine serum (FBS; cat. no. F7524, lot no. 022M3395, Sigma-Aldrich), 100 U/mL penicillin and 100 µg/mL streptomycin (penicillin-streptomycin solution 100x; cat. no. L0022, Biowest, Nuaillé, France). Alpha TC-1 cells (alpha TC-1 Clone 9, ATCC: CRL-2350) were cultivated in DMEM supplemented with 10% FBS, 15 mM HEPES, 0.1 mM non-essential amino acids, and 0.02% BSA. Cells were grown at 37°C, in humidified air containing 5% CO2. For transfection, 35,000 cells were seeded per cm2 of the cell culture dish, and incubated for 24 h. Then, they were incubated for another 24 h with a 1:3 DNA:PEI (Polyethylenimine MAX; MW 40,000, cat. no. 24765-2; Polysciences Inc., Warrington, PA, USA) solution containing 1.5 µg DNA per cm2 of transfected cells.

Page 51 of 200

Production of stably transgenic cell lines. Electroβ was produced via the following three-step procedure. (i) First, a 1.1E7-derived cell clone deficient in glucose-sensitive insulin secretion was selected. The resulting cell line, INSVesc, was used for step two. (ii) Second,

INSVesc was transduced with pProinsulin-NanoLuc-derived lentiviral particles, selected in culture medium containing 5 µg/mL blasticidin for 14 days, and cloned by limiting dilution. Monoclonal cell populations were tested for depolarization-triggered NanoLuc secretion. The monoclonal cell line showing the highest induction profile was used for step three. (iii) The proinsulin-NanoLuc-transgenic INSVesc was cotransfected with the SB100X expression vector pCMV-T7-SB100 (PhCMV-SB100X-pA) and the SB100X-specific transposon pKK66 (ITR-

PhEF1α-α1C-P2A-Kir2.1-pA-ITR:PRPBSA-BFP-P2A-PuroR-pA-ITR) and pMX251 (ITR-PhEF1α-

α2/δ1-P2A-β3-pA:PRPBSA-dTomato-P2A-BlastR-pA-ITR) as described above. After seven days, stably transgenic puromycin-resistant cell populations were selected by FACS-mediated single- cell sorting based on the cell’s fluorescence relative to parental 1.1E7 (BFP: 405 nm laser, 450/50 bandpass filter; dTomato: 561 nm laser, 570 nm long-pass filter, 586/15 bandpass filter) using a Becton Dickinson LSRII Fortessa flow cytometer (Becton Dickinson, Allschwil, Switzerland).

Lentiviral particle production. pProinsulin-NanoLuc-derived lentiviral particles were produced by cotransfecting HEK-293T cells with the lentiviral expression vector pProinsulin-

NanoLuc (PhCMV-Proinsulin-NanoLuc-pA), as well as the packaging (pSPAX2) and envelope (pMD2.G) vectors at a 4:3:1 ratio, respectively. After 48 h, the lentiviral particles were harvested from the cell culture supernatant and purified by 0.4 µm filtration.

Electrostimulation. (a) C-Dish. cells were electrostimulated using the C-Dish (Ionoptix, Dublin, Ireland), powered by an HP3245A Universal Source function generator (cat. no. 3245A; Hewlett Packard, Palo Alto, CA, USA) connected to a general-purpose linear amplifier (P200, cat. no. P200; FLC Electronics AB, Molndal, Sweden). C-Dish precisely places pairs of carbon electrodes into each well of a standard 6-well cell culture plate containing 35,000 cells/cm2 in 1.4 mL of cell-specific culture medium. (b) Cell culture chamber. Cell culture chambers with a 0.4 µm transparent PET membrane supporting the growth of monolayer cultures (Falcon® Permeable Support; cat. no 353095, Corning, Corning, NY, USA), fitting individual wells of a standard 24-well plate, were equipped with platinum electrodes (0.5 mm diameter; cat. no. HXA 050, Cooksongold Ltd., Birmingham, United Kingdom) placed above and below the cell-containing membrane and powered by the amplifier-connected HP3245A Universal Source function generator. The PET membrane of the cell culture chamber was

Page 52 of 200 seeded with 35,000 cells/cm2 cultivated in the cell-specific medium (400 µL inside the cell culture chamber, 1.3 mL inside the 24-well plate).

Analytical assays. SEAP (human placental secreted alkaline phosphatase) levels were profiled in cell culture supernatants using a colorimetric assay. 100 µL 2x SEAP assay buffer (20 mM homoarginine, 1 mM MgCl2, 21% diethanolamine, pH 9.8) was mixed with 80 µL heat-inactivated (30 min at 65°C) cell culture supernatant. After the addition of 20 µL substrate solution (120 mM p-nitrophenyl phosphate; cat. no. AC128860100, Thermo Fisher Scientific), the absorbance time course was recorded for 45 min at 405 nm and 37°C using a Tecan Genios PRO plate reader (cat. no. P97084; Tecan Group AG, Maennedorf, Switzerland) and the SEAP levels were determined as described previously141. Blood SEAP levels were quantified using a chemiluminescence-based assay (cat. no. 11779842001; Roche Diagnostics GmbH, Mannheim, Germany). In brief, serum was isolated from blood samples by centrifugation in Microtainer® serum separation tubes (SSTTM, cat. no. 365967; Becton Dickinson) for 5 min at 10,000×g. 50 µL of serum was heat-inactivated for 30 min at 65°C, centrifuged for 30 s at 5,000×g, transferred to a well of a 96-well plate containing 50 mL inactivation buffer, and incubated for 10 min at 20°C. 50 µL of substrate reagent was then added to each well and the sample was incubated for 10 min at 20°C. The luminescence was quantified using a Tecan Genios PRO plate reader (Tecan Group AG, Maennedorf, Switzerland). NanoLuc® luciferase was quantified in cell culture supernatants using the Nano-Glo® Luciferase Assay System (cat. no. N1110; Promega, Duebendorf, Switzerland). In brief, 7.5 µL of cell culture supernatant was added per well of a black 384-well plate and mixed with 7.5 µL substrate-containing assay buffer. Total luminescence was quantified using a Tecan Genios PRO plate reader (Tecan Group AG). Serum NanoLuc® was quantified on an EnVision 2104 Multilabel Reader (PerkinElmer, Waltham, Massachusetts, USA) using the ultrasensitive luminescence program. 5 µL of serum was diluted in 10 µL of ddH2O per well of a black 96-well plate and mixed with 15 µL of substrate- containing assay buffer. Aliquots of 15 µL of whole-blood samples were diluted in 5 μL of 50 mM EDTA and frozen at -20 °C until NanoLuc® quantification as described above.

Cell encapsulation. HEK-β cells were encapsulated in coherent alginate-poly-(L- lysine)-alginate beads (400 μm; 500 cells or 1-10 IEQs per capsule) using an Inotech Encapsulator Research Unit IE-50R (EncapBioSystems Inc., Greifensee, Switzerland) set to the following parameters: a 200-μm nozzle with a vibration frequency of 1025 Hz, a 25-mL syringe operated at a flow rate of 410 units, and a voltage of 1.12 kV for bead dispersion.

Page 53 of 200

Cell viability. Cell viability was quantified by incubating the cells for 2 h in working solution containing 60 µg/mL resazurin (cat. no. R7017, Sigma-Aldrich, Saint Louis, MO, USA), followed by fluorescence measurement of sample supernatants at 560/9 nm excitation and 590/20 nm emission using a Tecan Infinite 200Pro plate reader (Tecan Group AG, Maennedorf, Switzerland).

Bioelectronic implant for wireless electrostimulation in vivo. The casing of the bioelectronic implant containing the cell culture chamber and the electronics was 3D-printed using FDA-licensed polyamide and a Formiga P110 3D printer (EOS GmbH, Krailling, Germany; Fabb-It 3D Druckservice, Loerrach, Germany). To fit this casing, the height of the cell culture chamber (Falcon® Permeable Support; Corning, cat. no 353095 or Corning Transwell®; Corning, cat. no. CLS3462-48EA) was cut to 3 mm. The cell culture chamber’s 0.4 µm membrane enables nutrient supply and product delivery via the vascularized fibrous connective tissue of the host while protecting the electrosensitive designer cells from host-cell responses158,159. The switchboard, wireless electronics and pulse generator were designed and prototyped by Peter Buchmann, and the final printed circuit board was custom-manufactured (ITEAD Intelligent Systems Co. Ltd, Shenzhen, People’s Republic of China) (see Supplementary Information). The electrode above the cell culture chamber’s membrane containing the electrosensitive designer cells consists of a platinum wire (0.5 mm diameter; cat. no. HXA 050, Cooksongold Ltd.) and the electrode below the chamber’s membrane was made of a stainless-steel mesh (cat. no. 165; wire diameter 50 µm, aperture 100 µm; BOPP AG, Zurich, Switzerland) covering the entire membrane to provide a Faraday-cage effect that would protect the host from any electric shock. The bioelectronic implant was assembled by fitting the electronics into the casing, connecting the electrodes to the electronics, covering the electronic switchboard with insulating tape and sealing it with Epo-Tek 301-2 (cat. no. 301-2; Epoxy Technology Inc., Billerica, MA, USA), fitting the cell culture chamber into the casing, and positioning the electrodes. Bioelectronic implants were sterilized by immersion in 70% ethanol followed by rinsing in PBS. Each implant was completely filled with 300 µL of culture medium 6 6 containing 2 x 10 Electroβ suspension cells and 3 x 10 Electroβ cell aggregates, prepared by cultivation in AggreWell™400 (cat. no. 34460, STEMCELL Technologies GmbH, Koeln, Germany) via a dedicated hole, then sealed with PDMS-filled pipette tips (Sylgard 184, cat. no. 184.0001; Suter Kunststoffe AG, Fraubrunnen, Switzerland) and subcutaneously implanted into mice after 12 h. To adapt the bioelectronic implant for long-term repetitive in-situ exchange ® of individual Electroβ-cell batches (Fig. S20), we connected fill-in and exhaust tubes (Venofix Safety, B. Braun AG, Melsungen, Germany) to the cell culture chamber (Fig. S20A). After Page 54 of 200

6 implantation, 2 x 10 Electroβ cells were filled into the culture chamber. For cell replacement, a syringe was connected to the fill-in tube and the cell chamber was flushed several times with a physiological salt solution (0.9% NaCl) before the new batches of Electroβ cells were filled in. Fill-in tubes were closed by thermal welding.

Electronic circuitry of the field generator and the bioelectronic implant. The field generator provides wireless power transmission across living tissues using transmitter coils (L2 and L3) producing an alternating electromagnetic field that is detected by the receiver coil (L1) of the bioelectronic implant and converted into electric current to power and control all electronic components of the implant. The field generator also enables detection and functional monitoring of the bioelectronic implant. The field generator consists of four functional elements: the frequency generator, the resonance unit, the filter and edge shaper and the timer (Fig. S7, S8). (a) Frequency generator (Fig. S7A). A crystal-based oscillating circuit (IC2) connected to a negative impedance converter (T4, T6) drives the output transistor (T5) of the frequency generator. (b) Resonance unit (Fig. S7B). The output transistor controls the main transmitter coil (L3) current, which generates an alternating magnetic field. The secondary transmitter coil (L2) with an adjustable capacitor (C8) in parallel forms an oscillating circuit, which amplifies the resonance of the main transmitter coil (L3). (c) Filter and edge shaper (Fig S7C). The current drawn by the main transmitter coil (L3) generates a proportional voltage drop by passing a resistor (R1). This voltage drop is filtered out by an LC network (L1, C4, C5) and amplified by a class A amplifier (T3). A decoupling capacitor (C1) forms a high-pass filter and drives a positive-negative-positive (PNP) transistor (T1). The high gain factor of the PNP transistor produces digital-like behavior of the data output (Data Out) for each fast change of the main transmitter coil current and enables wireless communication between the field generator and the bioelectronics implant. (d) Timer (Fig. S5D). The data output (Data Out) triggers a timer circuit (IC1) that controls an LED, which enables functional monitoring of the bioelectronic implant in real time.

The electronic circuit of the bioelectronic implant consists of three functional elements: The power supply, the time generator and the communication circuit (Fig. S5, Fig. S6). (a) Power supply (Fig. S5A). The receiver coil (L1) and capacitors (C6, C8) connected in parallel form an oscillating circuit that is in resonance with the magnetic field of the transmitter coils of the field generator. To prevent potential impact of the electrical load on the resonance, power is withdrawn from two turns of the receiver coil via current transformation and rectified by two diodes (D4). (b) Time generator (Fig. S5B). Pulses are generated by an integrated time

Page 55 of 200 generator. The pulse length and intervals are set by external resistors (R7, R9), a diode (D11) and a capacitor (C14). The timer (IC4) includes an internal voltage divider with three resistors. The potential of the external capacitor and the voltage divider are compared by comparators and used to control the timer output. ON/OFF-states are stored by a flip-flop until the opposite voltage threshold is reached. One output of the flip-flop serves to discharge the capacitor (C14). To reach the required ratio between pulse duration and pulse interval of approximately 1:500, the capacitor is charged through a low-Ohmic resistor (R7) and a diode (D11) and discharged via a high-Ohmic resistor (R9). (c) Communication circuit (Fig. S5C). Upon detection of a negative pulse edge by a capacitor (C12) and a transistor (T3), the oscillating circuit of the power supply is short-circuited for 15-25 µs. This brief overload leads to a breakdown of the resonance in the receiver coil (L1); this results in an abrupt drop of the excitation current in the resonance unit of the field generator and produces a voltage increase. This voltage increase is filtered out by the excitation electronics in the field generator and provides digital information on the presence of the bioelectronic implant and the function of the time generator. All electronic components of the bioelectronic implant are listed in Table S8.

Animal experiments. To establish type-1 diabetes, wild-type 6-week old male Swiss mice (weighing 30-32 g RjOrl:SWISS; Janvier Labs, Le Genest-Saint-Isle, France) were fasted 18 h per day for two consecutive days and injected with a single dose of freshly diluted alloxan monohydrate (ALX; Sigma-Aldrich; cat. no. A7413, 195 mg/kg in 300 µL PBS) 46. Persistent fasting hyperglycemia was confirmed after 48 h using a glucometer (Contour®Next, Bayer Healthcare, Leverkusen, Germany). (a) Monitoring of study animals. General well-being of the animals was routinely monitored by animal caretakers by daily visual inspections. Project- specific monitoring was carried out at least five times per week by animal facility staff and animals were euthanized if symptoms of pain and/or distress were observed. Absence of the following applied humane endpoints was defined as asymptomatic disease state: weight loss > 20 %, insatiable polydipsia, recurrent dehydration (assessed by skin fold testing), surgical wound complications such as bleedings or inflammation/infection, abnormal breathing pattern, apathy or immobility, closed eyes or self-mutilation, detachment from the group or disinterest to the environment, un-kept appearance e.g. tangled fur and discoloration due to secretions. (b) Bioelectronic implant. At 72 h after alloxan treatment, the animals received inhalational isoflurane anesthesia and the bioelectronic devices were subcutaneously implanted on the back, with the polyester gauze-protected mesh electrode facing the ventral side. (c) Glucose tolerance test. Treated animals were fasted for 6 hours and then placed for 60 min below a field generator (see Supplementary Information) to power up the bioelectronic implant and Page 56 of 200 wirelessly stimulate insulin release by Electroβ cells immediately before the animals were subjected to glucose tolerance tests: blood glucose levels were monitored for indicated period after oral injection of 2 g/kg glucose. Mock-stimulated implants produced no insulin. (d) Human pancreatic islets implant. Human pancreatic islets were provided by the Geneva Islet transplantation center through the basic research program of European Consortium for Islet Transplantation (ECIT) supported by the Juvenile Diabetes Research foundation (JDRF; grant no 31-2008-416). At 72 h after alloxan treatment, the animals received anesthesia and the islets- containing chamber (2000 islet equivalents (IEQ) per implant) of a bioelectronic implant was placed as described above. (e) Real-time glycemia. Animals were fasted for 6 hours and glycemia was monitored for the indicated periods of time before and during electrostimulation. (f) In vivo kinetics. Implants were electrostimulated for 30 min and reporter protein levels were quantified in 15 μL blood samples, collected from the tail vein at the indicated time points. (g)

Biocompatibility. Mice were subcutaneously implanted on the back with either Electroβ-cell- containing implants, cell-free negative control implants, or biocompatible reference implants serving as positive controls according to ISO 10993. The biocompatible reference implants of identical shape were cast from polydimethylsiloxane (Sylgard 184, cat. no. 184.0001; Suter Kunststoffe AG, Fraubrunnen, Switzerland). They were autoclaved for 20 minutes at 121 °C prior to implantation. After three weeks, the animals were sacrificed and the bioelectronic implants as well as tissue samples were fixed in 10 % neutral buffered formalin and transferred to AnaPath GmbH (Oberbuchsiten/Liestal, Switzerland), where the biocompatibility assays were performed according to ISO 10993.

Animal experiments were performed according to the directive of the European Community Council (2010/63/EU), approved by the French Republic, and carried out by (i) Ghislaine Charpin-El Hamri (no. 69266309), Marie Daoud-El Baba (no. 69266310) at the Institut Universitaire de Technologie of the Université Claude Bernard Lyon 1, F-69622, Villeurbanne Cedex, France (project no. APAFIS # 16753 – CEEA-55 DR 2018-40v5), (ii) Marie-Didiée Hussherr and Shuai Xue (license number: 2996/30779) at the ETH Zurich in Basel, Switzerland, and by (iii) Jiawei Shao and Shuai Xue (license number: m20190102) in the laboratory of Haifeng Ye at the Institute of Biomedical Sciences and School of Life Sciences, East China Normal University, Shanghai, People’s Republic of China.

Histology. (a) Limited Systemic Toxicity. Two kidneys kidneys and two lobes from the liver of each mouse were processed by paraffin embedding, cut into 2-4 µm slices, and stained with hematoxylin and eosin (H&E) (Fig. S16). (b) Tissue samples. A small portion of

Page 57 of 200 the tissue in contact with the semipermeable membrane of a cell chamber was excised, embedded in paraffin, cut into 2-4 µm slices, and stained with H&E (Fig. S16). (c) Bioelectronic implant sections. Explanted bioelectronic devices, including surrounding tissue, were embedded in methyl methacrylate resin. The cell chamber of each implant was cut into 400 µm slices at its central position in the transverse direction using a diamond saw (EXAKT 300 CP System; EXAKT Technologies Inc., Oklahoma City, OK, USA), as shown in Fig. S16. These slices were ground to a thickness of 40-60 using the EXAKT 400 CS System (EXAKT Technologies Inc., Oklahoma City, OK, USA) and stained with Paragon (toluidine blue and basic fuchsin). The same procedure was applied to obtain sections at the switchboard’s central position (Fig. S16).

Immunohistochemistry (IHC). IHC was performed using BOND-III Fully Automated IHC and ISH Stainer (Leica Biosystems, Wetzlar, Germany). Rabbit monoclonal anti-CD11b (Ab133357; dilution 1:5000, lot no. GR3209213-2; Abcam, Cambridge, United Kingdom) and rabbit monoclonal anti-CD68 (Ab125212; dilution 1:1000, lot no. GR300618-28; Abcam, Cambridge, United Kingdom) were used. Prior to staining, the samples were incubated in Bon Epitope Retrieval Solution 1 (citrate-based buffer, pH: 5.9-6.1; Leica Biosystems, Wetzlar, Germany) at 100 °C for 20 min.

Histopathology. Histopathology of the implantation site was analysed at AnaPath GmbH using a scoring system according to ISO 10993-6:2016(E). Representative images were taken with an Olympus UC30 camera.

Image Analysis. Quantitative analyses of immunohistochemistry sections were performed using the image analysis software QuPath (https://qupath.github.io).

Cytokine Profiling. Serum samples of each treatment group were mixed and analyzed with a Proteome Profiler Array, Mouse Cytokine Array Panel A (cat. no. ARY 006; R&D Systems, Inc., Minneapolis, USA), according to the manufacturer’s instructions. Chemiluminescence was analysed using an ImageQuant LAS 4000 mini (GE Healthcare Life Sciences, Chicago, Illinois, USA).

Hematology. Hematological analysis was performed using the scil Vet abc Plus+ analyzer (scil animal care company GmbH, Viernheim, Germany).

Sample size determination and statistics. Data are presented as mean values. Error bars show the standard error of the mean. The “n” number refers to biological replicates. Sample size for in vivo experiments was chosen to provide statistical power (1-β) ≥ 0.8 and type I error

Page 58 of 200 rate (α) ≤ 0.05 for a 30 % change of the mean in treated groups, assuming 25 % standard deviation. The p-value was calculated by performing the two-tailed t-test. Outliers were considered by applying the ROUT method180.

Page 59 of 200

Supplementary Figures

Fig. S1 Validation of system components.

HEK293T cells were co-transfected with plasmids encoding the indicated channel subunits (+): pCaV1.2 (PhCMV-α1C-pA), pCaVα2δ1 (PhCMV-α2/δ1-pA) and pCaVβ3 (PhCMV-β3-pA), pKir2.1

(PhCMV-Kir2.1-pA) or mock DNA (-), as well as PNFAT3-driven SEAP expression plasmid (pMX57). Cell membrane was depolarized with 40 mM KCl for one hour (red bars) and after 24 hours SEAP production was quantified in supernatant samples. Blue bars represent negative control samples. ns – not significant, ***p<0.001.

Page 60 of 200

Fig. S2 Impact of salt concentration on the electrogenetic circuit.

(A) ElectroHEK cells were cultivated for 24 hours in medium supplemented with various concentrations of sodium chloride (NaCl), sodium phosphate (Na2PO4), magnesium sulphate

(MgSO4), calcium chloride (CaCl2), or potassium chloride (KCl), and then SEAP was quantified in the culture supernatants. (B) Electroβ cells were cultivated for 10 min in medium containing various concentrations of NaCl, Na2PO4, MgSO4, CaCl2 or KCl, and then NanoLuc was quantified in the culture supernatants. Data points represent mean ± SEM. Statistical significance was calculated versus time 0. ns – not significant, * p<0.05, **p<0.01, ***p<0.001.

Page 61 of 200

Page 62 of 200

Fig. S3 Electrically inducible gene expression system.

(A-D) Toxicity of electrical stimulation. HEK293T cells were co-transfected with plasmids encoding the L-type voltage-gated calcium channel CaV1.2: pCaV1.2, (PhCMV-α1C-pA), pCaVα2δ1 (PhCMV-α2/δ1-pA), pCaVβ3 (PhCMV-β3-pA), Kir2.1: pKir2.1 (PhCMV-Kir2.1-pA). Electrical stimulation (orange bars) was applied at the indicated voltage (A,C), or for the indicated period of time (B,D), and 24 hours later cell viability was assessed. (E) Electrical stimulation of cells lacking a voltage-gated calcium channel. HEK293T cells were co- transfected with pCDNA3.1(+) and PNFAT3-driven SEAP expression plasmid (pMX57). Electrical stimulation (50 V, 1 Hz, 1 hour and pulse length as indicated below the graph; orange bars) was applied and 24 hours later SEAP was quantified in supernatant samples. Depolarization with 40 mM KCl for 1 hour was used as a positive control (red bar). Blue bars represent unstimulated control. Bars represent mean ± SEM. n = 3. ns – not significant, * p<0.05, **p<0.01.

Page 63 of 200

Fig. S4 Electrostimulation kinetics.

HEK293T cells were co-transfected with plasmids encoding pCaV1.2 (PhCMV-α1C-pA), pCaVα2δ1 (PhCMV-α2/δ1-pA) and pCaVβ3 (PhCMV-β3-pA), pKir2.1 (PhCMV-Kir2.1-pA) and PNFAT3- driven SEAP and mINS expression plasmid (pKK137). (A) SEAP production kinetics. Cells were electrostimulated with 50 V, 2 ms pulses at 1 Hz frequency for one hour (red line). SEAP concentration was quantified in supernatant samples every one hour. Blue line represents control samples (not stimulated). (B) Reversibility assay. Cells were electrostimulated at the 0 and 24 h time points (blue line), or at the 12 h time point (red line). The culture medium was exchanged every 12 hours and the electrostimulation status was changed (from ON to OFF or from OFF to ON). SEAP was quantified in supernatant samples. Data points represent mean ± SEM. n = 3. Statistical significance was calculated between on (red dots) and off (blue dots) states for a given time point. * p<0.05, **p<0.01, ***p<0.001.

Page 64 of 200

Page 65 of 200

Fig. S5 Electronic circuit of the bioelectronic implant.

The three main functional elements are marked with dashed frames: the power supply (A), the time generator (B) and the communication circuit (C). (A) Power supply. The coil (L1) and capacitors (C6, C8) connected in parallel form an oscillating circuit that is in resonance with the magnetic field of the transmitter coil. To prevent potential impact of the electrical load on the resonance, power is withdrawn from two turns of the receiver coil via current transformation and rectified by two diodes (D4). (B) Time generator. Pulses are generated by an integrated time generator. Pulse length and intervals are set by external resistors (R7, R9), a diode (D11) and a capacitor (C14). The timer (IC4) includes an internal voltage divider with three resistors. The potential of the external capacitor and the voltage divider are compared by comparators and used to control the timer output. ON/OFF-states are stored by a flip-flop until the opposite voltage threshold is reached. One output of the flip-flop serves to discharge the external capacitor (C14). To reach the required ratio between pulse duration and pulse interval of approximately 1:500, the capacitor is charged through a low-Ohmic resistor (R7) and a diode (D11) and discharged via a high-Ohmic resistor (R9). (C) Communication circuit. Upon detection of a negative pulse edge by a capacitor (C12) and a transistor (T3), the oscillating circuit of the power supply is short-circuited for 15-25 µs. This brief overload leads to a breakdown of the resonance in the receiver coil (L1) and results in an abrupt drop of the excitation current in the transmitter coil and produces a voltage increase.

Page 66 of 200

Fig. S6 Bioelectronic implant switchboard.

(A) Top view of the bioelectronic implant switchboard. Dashed frames correspond to functional elements described in Figure S5: power supply (yellow), time generator (white) and communication circuit (red). (B) Schematic view of the bioelectronics implant. Conductive paths are marked with lighter green. Components are marked with white frames. Orange rectangles represent connectors. Components are described in Table S8.

Page 67 of 200

Page 68 of 200

Fig. S7 Electronic circuit of the field generator.

Four main functional elements are marked with dashed frames: the frequency generator (A), the resonance coils (B), the filter and edge shaper (C), and the timer (D). (A) Frequency generator. A crystal-based oscillating circuit (IC2) connected to a negative impedance converter (T4, T6) drives the output transistor (T5) of the frequency generator. (B) Resonance coils. The output transistor controls the coil (L3) current, which generates an alternating magnetic field. The second coil (L2) with an adjustable capacitor (C8) in parallel forms an oscillating circuit, which amplifies resonance of the main coil. (C) Filter and edge shaper. The current drawn by the resonance coil generates a proportional voltage drop by passing a resistor (R1). This voltage drop is filtered out by an LC network (L1, C4, C5) and amplified by a class A amplifier (T3). A decoupling capacitor (C1) forms a high-pass filter and drives a PNP transistor (T1). The high gain factor of the PNP transistor produces digital-like behavior of the data output (Data Out) for each fast change of the coil current and enables wireless communication from the implant to the field generator. (D) Timer. The data output triggers a timer circuit (IC1) that drives an LED, which enables function monitoring of the bioelectronics implant.

Page 69 of 200

Page 70 of 200

Fig. S8 Field generator switchboard.

(A) Schematic view of the field generator. Conductive paths are marked with lighter green. Components are marked with white frames. Orange rectangles represent connectors. (B) Enlarged view of the part enclosed with a black frame in (A). Dashed frames correspond to the functional elements described in Figure S7. Frequency generator (yellow), the resonance coils (purple), the filter and edge shaper (red), and the timer (white).

Page 71 of 200

Fig. S9 Position-dependent coupling intensity between the field generator and the bioelectronic device.

(A) Schematic representation of the experimental design. The Y-axis indicates the vertical position of the implant. The X-axis indicates the horizontal position of the implant measured from the middle point. (B) Dependence of the impulse voltage reached in the bioelectronic implant upon the horizontal (0 – 8 cm) and vertical (0 – 4 cm) positions.

Page 72 of 200

Fig. S10 Long-term electro-stimulation stability of the Electroβ cell line.

5 2 x 10 Electroβ cells were either electro-stimulated for 20 minutes (orange bars) using 2 ms pulses at 5 Hz and 10 V or not electro-stimulated (blue bars) on day 1 and after a continuous cultivation period of 90 days. NanoLuc levels were quantified in the culture supernatants. Bars represent mean ± SEM. n = 3. ns – not significant, **p<0.01, ***p<0.001.

Page 73 of 200

Fig. S11 Comparative insulin-release kinetics.

Comparative insulin-release kinetics of Electro (constitutive expression, vesicular secretion) and

ElectroHEK, OptoHEK and HEK- (PNFAT-driven expression, secretory pathway). (A) ElectroHEK

(4 hours electrostimulation), (B) OptoHEK (24 hours blue-light illumination) and (C) HEK- (40 mM glucose over 24 hours) were stimulated (green line) or non-stimulated (blue line) and insulin secretion was profiled in the culture supernatant. Likewise, (D) Electroβ were (15 minutes electrostimulation) stimulated (green line) or non-stimulated (blue line) and the immediate insulin release was monitored in the culture supernatant.

Page 74 of 200

Fig. S12 Electrically inducible fast insulin secretion system.

(A, B) Voltage dependence. Electroβ cells were stimulated with 2 ms pulses at 1 Hz frequency and at the indicated voltage. NanoLuc was quantified from supernatant samples before (blue bars) and after electrical stimulation (orange bars) from above (A) and below (B) the cell layer. (C, D) Pulse length dependence. Electroβ cells were stimulated with 7.5 V pulses at 5 Hz frequency with the indicated pulse length. NanoLuc was quantified from supernatant samples before (blue bars) and after electrical stimulation (orange bars) from above (C) and below (D) the cell layer. The statistical significance of differences between the negative control and each of the stimulation conditions was calculated. Bars represent mean ± SEM. n = 3. ***p<0.001, ****p<0.0001.

Page 75 of 200

Fig. S13 Electro-stimulated glucagon secretion by pancreatic -cells.

2 x 105 pancreatic -cells (α-TC1) cells were seeded into cell-culture inserts and electro-stimulated with 2 ms pulses at 5 Hz and 10 V for 20 minutes (red bar). Then, glucagon levels were quantified in the culture supernatant. Bars represent mean ± SEM. n = 3. *p<0.05.

Page 76 of 200

Fig. S14 In vivo kinetics of protein secretion.

(A) HEK-β cell line secretion kinetics in vivo. HEK-β cells were encapsulated into alginate beads and injected into hyperglycemic mice. Insulin was quantified from serum samples after 4, 24 and 48 hours. Data points represent mean ± SEM. n = 7. (B) Electroβ secretion kinetics in vivo. Electroβ cells inside the bioelectronic implant were electrostimulated for 30 minutes. Insulin levels in whole-blood samples were quantified before the stimulation, and after 30, 60, 120 minutes and 24 hours.

Page 77 of 200

Fig. S15 Impact of materials used in the biomedical device on cell viability.

The viability of parental HEK-293T cells was measured after incubation for 24 hours with implant materials: solidified PDMS (biocompatible reference material), stainless steel (electrode material) and Epo-Tek 301-2 (implant material). Bars represent mean ± SEM. n = 3. Statistical significance was calculated between cells only (negative control) and each of the implant materials. ns – not significant, *p<0.05.

Page 78 of 200

Page 79 of 200

Fig. S16 Biocompatibility.

Histological sections of tissue-surrounded bioelectronic implant. Devices containing

Electroβ cells were surgically placed on the dorsal side of mice for 21 days. Fibrotic capsule formation was examined by Paragon staining. (A) Cross section of the cell- containing chamber. Green arrow, porous membrane; red arrows, 3D-printed frame; orange arrows, electronic switchboard; blue arrows, cell insert frame; fibrotic tissue, yellow arrows (please note that, compared to the structural part of the biocompatible implant material of the bioelectronic device (Fig. S15), the fibrotic tissue formation at the porous membrane where the Electro cells interface with the host tissue, is less prominent. (B) Enlarged image of the fibrotic tissue formed at the implant material of the bioelectronic device. Red arrows - blood vessels, green arrows – fibroblasts, blue arrows – fibrocytes, yellow arrows – granulocytes, orange arrows - lymphocytes.

Page 80 of 200

Page 81 of 200

Fig. S17 Local immune response.

(A) Local effects on the immune response after implantation were scored according to ISO 10993. A higher score indicates greater local effects and lower biocompatibility. (B,C) Immunohistochemistry. Fragments of tissue adherent to the porous membrane of a bioelectronic implant were excised and immunostained. Graphs represent percentage of positively stained cells in experimental groups. (B) Anti-CD11b antibodies. CD11b is expressed on macrophages, neutrophils and natural killer cells. (C) Anti-CD68 antibodies. CD68 is expressed on activated macrophages. ns – not significant.

Page 82 of 200

Fig. S18 Dot-blot-based cytokine profiling of treated mice.

Reference inflammatory cytokines CXCL-10, CXCL-13 and C5a were profiled in the serum of mice 3 weeks after the introduction of bioelectronic implants with (A) and without (B) cells, or biocompatible reference implants (C).

Page 83 of 200

Fig. S19 Impact of bioelectronic implants on the body weight of mice.

The body weight of mice was profiled over three weeks after the introduction of bioelectronic implants with and without ElectroHEKs or biocompatible control implants. The weight of the implant (2 g) was subtracted from the body weight after implantation. Data points represent mean ± SEM. n = 8. There were no significant differences between the groups after implantation.

Page 84 of 200

Page 85 of 200

Fig. S20 Bioelectronic implants modified for long-term glycemic control in type- 1-diabetic mice.

(A) Picture of bioelectronic implants containing fill-in and exhaust tubes to allow repetitive in-situ exchange of Electroβ-cell batches. (B) Validation of implant functionality over 3 weeks in type-1-diabetic mice. Bioelectronic devices shown in (A) were implanted into type-1 diabetic mice (T1DM), allowing weekly exchange of the

Electroβ-cell batches. Insulin levels were quantified once a week. Data represent mean ± SEM, n = 5 mice. Statistical significance of blood-insulin levels was calculated against blood-insulin levels before implantation and against homeostatic insulin levels of wild- type mice. (C) Electro-stimulated insulin secretion kinetics of bioelectronic implants in treated type-1-diabetic mice after three weeks. Electroβ-cell-containing bioelectronic implants were electro-stimulated for 30 minutes and resulting blood-insulin levels were profiled. Data represent mean ± SEM, n = 5 mice per group. (D) Picture of representative bioelectronic devices explanted from mice after termination of the experiment described in (B).

Page 86 of 200

Supplementary Tables

Table S1. Temperature dependence of implant functionality.

Supply Voltage [V] 25°C 30°C 35°C 40°C 45°C 50°C

3 2.32 2.32 2.32 2.32 2.32 2.31

4 2.39 2.37 2.38 2.38 2.38 2.38

5 2.46 2.44 2.45 2.44 2.45 2.45

6 2.51 2.49 2.5 2.49 2.48 2.48

7 2.55 2.53 2.52 2.52 2.51 2.51

8 2.58 2.55 2.55 2.54 2.52 2.52

9 2.62 2.6 2.59 2.58 2.57 2.56

10 2.66 2.62 2.62 2.61 2.59 2.57

11 2.69 2.63 2.63 2.61 2.6 2.58

12 2.7 2.63 2.64 2.62 2.6 2.58

The implant was equipped with a 1 kOhm load at the impulse output, placed on a hot plate and covered with a 3 cm layer of polystyrene. The circuit was powered with an adjustable power supply. The output voltage was measured with a digital oscilloscope, LeCroy Wavesurfer, at the indicated temperature.

Page 87 of 200

Table S2. Quality control of bioelectronic devices

Water immersion test1

Batch size (no. of items) 50

Implants containing functional electronics after 24 h of water immersion 48

% Functional 96

Cell leakage test (mechanical integrity)2

Number of devices 5

Number of devices with cell leakage detected 0

Electronic functionality in vivo3

Number of devices 10

Functionality at week 1 after implantation 10

Functionality at week 2 after implantation 10

Functionality at week 3 after implantation 10

1A batch of 50 implants was placed in 20 cm deep water for 24 hours. Electronic functionality of the implants was tested after 24 hours.

2Bioelectronic devices were filled with 2 x 105 cells and incubated in 60 mm cell culture dishes. After 12 hours, the cell-containing bioelectronic devices were transferred to fresh cell culture dishes, incubated for 5 days and carefully examined to determine whether or not cells were present outside the bioelectronic device.

Page 88 of 200

3Functionality of bioelectronic devices used for experiment of Fig. S20B measured by a resonance-indicating diode. Please refer to Fig. 4 and Fig. S7 for details.

Page 89 of 200

Table S3. Systemic toxicity.

Animal Group Organ Findings Number Kidney Pyelitis, chronic, focal, unilateral, minimal 1 Liver Hemopoietic cell foci, minimal Kidney Pyelitis, chronic, focal, unilateral, minimal 5 Liver No findings Kidney No findings Biocompatible reference item 14 Liver Inflammatory cell foci, minimal Kidney Mononuclear cell foci, unilateral, minimal 18 Liver Hemopoietic cell foci, minimal Kidney Pyelitis, chronic, focal, biilateral, minimal 19 Liver No findings Kidney Mononuclear cell foci, unilateral, minimal 3 Liver No findings Kidney Mononuclear cell foci, unilateral, minimal 4 Liver Inflammatory cell foci, minimal Implant without cells Mononuclear cell foci, unilateral, minimal Kidney Pyelitis, chronic, focal, unilateral, minimal 12 Cast, hyaline, focal, unilateral, minimal Liver Hemopoietic cell foci, minimal 13 Kidney Pyelitis, chronic, focal, unilateral, minimal

Page 90 of 200

Animal Group Organ Findings Number Liver Hemopoietic cell foci, minimal Cyst, cortical, present Kidney Urothelial hyperplasia, unilateral, minimal 16 Pyelitis, chronic, focal, unilateral, minimal Liver No findings Kidney No findings 8 Liver Inflammatory cell foci, Slight Kidney Mononuclear cell foci, unilateral, minimal 9 Hemopoietic cell foci, minimal Liver Inflammatory cell foci, minimal Pyelitis, chronic, focal, unilateral, minimal Kidney Implant with cells 15 Cast, hyaline, focal, unilateral, minimal Liver Hemopoietic cell foci, minimal Kidney Pyelitis, chronic, focal, unilateral, minimal 21 Liver Hemopoietic cell foci, minimal Kidney Mononuclear cell foci, unilateral, minimal 22 Liver Hemopoietic cell foci, minimal

Page 91 of 200

Table S4. Hematology

Implant with

Biocompatible control implant cells Implant without cells

Mean Standard deviation Mean Standard deviation Mean Standard deviation

White blood cells [103/mm3] 11.08 3.33 13.18 4.17 13.90 3.76

Lymphocytes [103/mm3] 6.70 2.25 7.80 2.46 8.19 2.03

Monocytes [103/mm3] 0.55 0.18 0.73 0.32 0.71 0.22

Granulocytes [103/mm3] 3.83 0.99 4.65 1.82 5.00 2.04

Eosinophils [103/mm3] 1.43 0.65 1.09 0.64 1.06 0.78

Erythrocytes [106/mm3] 7.94 0.60 7.17 0.52 7.28 0.73

Blood samples were taken from mice 21 days after implantation of one of the three items. White blood cells, lymphocytes, monocytes, granulocytes, eosinophils and erythrocytes were counted. n = 6 for the biocompatible control implant group and n = 8 for the other two groups.

Page 92 of 200

Table S5. Biocompatibility - single animal findings according to adapted ISO 10993-6:2016(E) scoring system

Biocompatible reference Implant without cells Implant with cells

Replica of Chamber Section Chamber Section Chamber Section Switchboard Section

Animal: 1 5 14 18 19 3 4 12 13 16 8F 9 15 21 22 8 9 15 21 22

Polymorphonuclear 2 2 3 2 1 2 2 1 2 3 2 3 2 2 2 1 2 1 2 2

Lymphocytes 1 2 2 2 2 1 2 2 1 1 2 2 2 1 1 2 1 2 1 1

Plasma cells 0 0 0 0 1 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0

Macrophages 2 2 2 2 1 2 2 2 1 1 2 1 1 1 2 2 1 1 1 1

Giant cells 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0

Necrosis 0 0 0 0 0 1 0 0 0 0 0 0 0 0 0 0 0 0 0 0

Subtotal (x2) 10 12 14 12 10 12 12 10 8 10 12 12 10 8 10 10 8 8 8 8

Neovascularisation 2 2 3 2 2 1 1 1 1 1 2 1 2 2 1 1 1 1 1 2

Fibrosis 3 3 3 3 3 2 3 3 3 2 3 3 3 2 3 2 2 2 2 3

Fatty infiltrate 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0

Detritus within fibrous capsule 1 1 1 2 0 2 1 0 0 1 2 1 1 1 1 1 1 0 1 1

Subtotal 6 6 7 7 5 5 5 4 4 4 7 5 6 5 5 4 4 3 4 6

Total 16 18 21 19 15 17 17 14 12 14 19 17 16 13 15 14 12 11 12 14

Page 93 of 200

Table S6. Energy consumption of optogenetic and electrogenetic implants.

Peak current Average current Predicted theoretical Description Source [mA] [mA] battery lifetime [h]

Near infrared optogenetic implant 28 14 143 Folcher et al.17

Near infrared optogenetic implant “HydroLED” 350* 350 5.7 Shao et al.16

Electrogenetic implant 13 0.027 75000 This work

Average current was calculated assuming that no current is drawn between pulses. Predicted theoretical battery lifetime was calculated for a typical battery used for implantable cardiac pacemakers, which is 2 Ah175. The electrogenetic implant used in this work generates peak current of around 13 mA in 2 ms pulses at 1 Hz frequency. The optogenetic implant published by Folcher et al.17 generates peak current of up to 28 mA for 30 s followed by 30 s of resting time. *Authors did not measure the output current. The calculation was based on typical LED current value provided by the manufacturer.

Page 94 of 200

Table S7. Plasmids used and designed in this study

Plasmid Description and cloning strategy Reference

181 pCavα1C Constitutive α1C expression vector (PhCMV-α1C-pA). (Addgene no. 26572).

182 pCavα1D42 Constitutive α 1D42 expression vector (PhCMV-α1D42-pA). (Addgene no. 49332).

182 pCavα1D42A Constitutive α1D42A expression vector (PhCMV-α1D42A-pA). (Addgene no. 49333).

183 pCavα2δ1 Constitutive α2/δ1 expression vector (PhCMV-α2/δ1-pA). (Addgene no. 26575).

184 pCavβ3 Constitutive β3 expression vector (PhCMV-β3-pA). (Addgene no. 26574).

pcDNA3.1(+) Constitutive mammalian expression vector containing a NeoR resistance gene (PhCMV-MCS-pA) Thermo Fisher Scientific, CA

pcDNA3.1/ Constitutive mammalian expression vector containing a HygroR resistance gene (PhCMV-MCS-pA). Thermo Fisher Hygro(+) Scientific, CA

185 pCMV-T7-SB100 Constitutive SB100X expression vector (PhCMV-SB100X-pA). (Addgene no. 34879).

pEGFP-N1 Constitutive EGFP expression vector (PhCMV-EGFP-pA). Clontech, CA

Page 95 of 200

Plasmid Description and cloning strategy Reference pGEM-T Easy Bacterial expression vector. Promega, WI

186 pGEMTEZ-Kir2.1 pGEM-T Easy containing Kir2.1. (Addgene no. 32641).

187 pMD2.G Constitutive VSV-G expression vector (PhCMV-VSV-G-pA). (Addgene no. 12259).

188 pMS2-P65- Lentiviral vector containing a PhEF1α-driven expression unit. (Addgene no. 61426). HSF1_Hygro

46 pMX57 PNFAT3-driven SEAP expression vector (PNFAT3-SEAP-pA). pMX251 SB100X-specific transposon containing a constitutive dTomato and BlastR expression unit and a constitutive 46

expression unit for α2/δ1 and β3 (ITR-PhEF1α-α2/δ1-P2A-β3-pA:PRPBSA-dTomato-P2A-BlastR-pA-ITR).

168 pProinsulin-NanoLuc Lentiviral vector for constitutive expression of Proinsulin-NanoLuc (PhCMV-Proinsulin-NanoLuc-pA). (Addgene no. 62057).

189 pSBbi-BP SB100X-specific transposon containing a constitutive BFP and PuroR expression unit (ITR-PhEF1α-MCS-pA:PRPBSA- BFP-P2A-PuroR-pA-ITR). (Addgene no. 60512)

pSEAP2-Control Constitutive mammalian SEAP expression vector (PSV40-SEAP-pA) Clontech, CA psPAX2 Lentiviral packaging vector. (Addgene no. 12260) 190

Page 96 of 200

Plasmid Description and cloning strategy Reference

pKK5 Constitutive Kir2.1 expression vector (PhCMV-Kir2.1-pA). Kir2.1 was excised from pGEMTEZ-Kir2.1 using EcoRI This work and cloned into the corresponding site (EcoRI) of pcDNA3.1(+).

pKK55 Constitutive α1C and Kir2.1 expression vector (PhEF1α-α1C-P2A-Kir2.1-pA). This work

The vector backbone of pEGFP-N1 was PCR-amplified with OKK119 (5’-CTTA AGGCGAGAATCGGAGATATGATAAAGCGGCCGCGACTCTAG-3’) and OKK108 (5’-GCCCACT

GACGGGCACTAATAACTAATGCATGGCGGTAATACGGTTATC-3’), PhEF1α was PCR-amplified from pMS2-P65-HSF1_Hygro with OKK107 (5'-GATAACCGTATTACCGCCATGCATTAGTTATTA

GTGCCCGTCAGTGGGC-3') and OKK110 (5'-GGTGGCGAATTCAAGCTTGCTAGCGTCACGACA

CCTGAAATGGAAG-3'), α1C was PCR-amplified from pCavα1C with OKK109 (5'- GTGTCGTGACGCTAGCAAGC TTGAATTCGCCACCATGGTCAATGAAAACACGAGGATGTAC-3') and OKK112 (5'-CTCCTCCA CGTCTCCAGCCTGCTTCAGCAGGCTGAAGTTAGTAGCTCCGC

TTCCCAGGTTGCTGACGTAGGACC-3’) and Kir2.1 was PCR-amplified from pKK5 with OKK111 (5'- CTTCAGCCTGCTGAAG CAGGCTGGAGACGTGGAGGAGAACCCTGGACCTATGGGCAGTGTG

AGAACCAAC-3') and OKK120 (5' CTAGAGTCGCGGCCGCTTTATCATATCTCCGATTCTCGCCT

TAAG 3') and all fragments were assembled by Gibson cloning.

pKK56 Constitutive α2/δ1 and β3 expression vector (PhEF1α-α2/δ1-P2A-β3-pA). This work

Page 97 of 200

Plasmid Description and cloning strategy Reference

The vector backbone of pcDNA3.1/Hygro(+) was PCR-amplified with OKK130 (5'- GACAGCTACTGATGAGCGGCCGCTCGAGTCGATCAGCCTCGACTGTG-3') and OKK126 (5'-

GCCCACTGACGGGCACTAATCAACGCGTATATCTGGCCCGTAC-3'), PhEF1α was PCR-amplified from pMS2-P65-HSF1_Hygro with OKK125 (5'-GTACGGGCCAGATATACGCGTTGATTAGTGCCCGTCAGTGGGC-3') and OKK122 (5'- CAGGCAGCCAGCAGCCATGGTGGCGAATTCAAGCTTGCTAGCGTCACGACACCTGAAATGGAAG -

3'), α2δ1 was PCR-amplified from pCaVα2δ1 with OKK121 (5'-CTTCCATTTCAGGTGTCGTGACGCTAGCAA GCTTGAATTCGCCACCATGGCTGCTGGCTGCCTG-3') and OKK116 (5'-CTACGTCCCCGGCCT

GTTTGAGAAGGCTGAAGTTGGTCGCTCCGCTCCCCCATAGATAGTGTCTGCTGCCAG-3') and β3 was

PCR-amplified from pCaVβ3 with OKK115 (5'-CAACTTCAGCCTTCTCAAACAGGCCGGGGAC GTAGAGGAGAACCCCGGGCCGATGTATGACGACTCCTACGTGC-3') and OKK124 (5'- GACTCGAGCGGCCGCTCATCAGTAGCTGTCCTTAGGCCAAG-3') and all fragments were assembled by Gibson cloning. pKK66 SB100X-specific transposon containing a constitutive BFP and PuroR expression unit and a constitutive expression This work

unit for α1C and Kir2.1 (ITR-PhEF1α-α1C-P2A-Kir2.1-pA:RPBSA-BFP-P2A-PuroR-pA-ITR). α1C-P2A-Kir2.1 was PCR-amplified from pKK55 using OKK141 (5’-attaggcctctgaggccA TGGTCAATGAAAACACGAGGATGTACGTTC-3‘) and OKK142 (5’-attaggcctgacaggccTCATAT CTCCGATTCTCGCCTTAAGGGC-3‘), restricted with SfiI and cloned into corresponding site (SfiI) of pSBbi-BP.

Page 98 of 200

Plasmid Description and cloning strategy Reference

pKK137 PNFAT3-driven SEAP and mINS expression vector (PNFAT3- SEAP-P2A-mINS-pA). SEAP-P2A-mINS was excised This work from pMX256 using EcoRI/SalI, and cloned into corresponding sites of pMX57.

Oligonucleotides: Restriction endonuclease-specific sites are shown in italics and annealing sequences are shown in capital letters.

Abbreviations: 1C, 1C subunit of the murine L-type voltage-gated calcium channel Cav1.2; α1D42A, α1D42A subunit of the murine L-type voltage-gated calcium channel Cav1.3; 1D42, 1D42 subunit of the murine L-type voltage-gated channel Cav1.3; α2/δ1, α2 and δ1 subunits of the murine L-type voltage-gated calcium channel Cav1.2; 3, 3 subunit of the murine L-type voltage-gated calcium channel Cav1.2; BFP, blue fluorescent protein; BlastR, blasticidin resistance gene; Cav1.2, member 2 of the Cav1 family of L-type voltage-gated calcium channels; Cav1.3, member 3 of the Cav1 family of L-type voltage-gated calcium channels; dTomato, dimeric red fluorescent protein variant; EGFP, enhanced green fluorescent protein; HygroR, hygromycin resistance gene; ITR, inverted terminal repeats of SB100X; Kir2.1, murine inwardly rectifying potassium channel; MCS, multiple cloning site; mINS, modified insulin variant for optimal expression in HEK-293 cells ; NanoLuc, Oplophorus gracilirostris luciferase; NeoR, neomycin resistance gene; NFAT, nuclear factor of activated T-cells; pA, polyadenylation signal; P2A; porcine teschovirus-1 2A self-cleaving peptide; PCR, polymerase chain reaction; PhCMV, human cytomegalovirus immediate early promoter; PhEF1α, human elongation factor 1 alpha promoter; PNFAT3, synthetic mammalian promoter containing five tandem repeats of a human IL-4 NFAT-binding site; Proinsulin-NanoLuc, modified proinsulin with its C-peptide replaced by NanoLuc; PRPBSA, constitutive synthetic mammalian promoter;

PSV40, simian virus 40 promoter; PuroR, puromycin resistance gene; SB100X, optimized Sleeping Beauty transposase; SEAP, human placental secreted alkaline phosphatase; VSV-G, vesicular stomatitis virus protein G; ZeoR, zeocin resistance gene.

Page 99 of 200

Table S8. Electronic components of the bioelectronic implant

Designator Description Manufacturer Manufacturer part number BAS28 High-speed double diode NXP BAS28,215 BC856 PNP Epitaxial Silicon Transistor NXP BC856B ZD15 Voltage regulator diodes BZX84 NXP BZX84/C15 BAT54C Schottky barrier (double) diode NXP BAT54C LMC555 CMOS Timer Texas Instruments LMC555CMM/NOPB 10 k Resistor 10 k Vishay CRCW080510K0FKEA 1.5 k Resistor 1.5 k Vishay CRCW08051K50FKEA 1.6 M Resistor 1.6 M Vishay CRCW08051M60FKEA 100 pF Capacitor 100 pF Murata GRM2165C1H101JA01D 68 pF Capacitor 68 pF RND Electronics RND1500805N680J500 1 uF Capacitor 1 uF RND Electronics RND1500805B105K250 6.5-30 pF Trimmer capacitor Murata TZC3P300A310R00 Manufacturers: NXP Semiconductors, Eindhoven, Netherlands; Texas Instruments, Dallas, TX, USA; Vishay, Malvern, PA, USA; RND Electronics, Nänikon, Switzerland; Murata Manufacturing Co., Nagaokakyō, Japan.

Page 100 of 200

CHAPTER II: Rewiring of endogenous signalling pathways to genomic targets for therapeutic cell reprogramming

Krzysztof Krawczyk1, Leo Scheller1, Hyojin Kim1 and Martin Fussenegger1,2,*

1Department of Biosystems Science and Engineering, ETH Zurich, Mattenstrasse 26, CH-4058 Basel, Switzerland.

2Faculty of Science, University of Basel, Mattenstrasse 26, CH-4058 Basel, Switzerland

*Corresponding author. E-mail: [email protected]

Nature Communications volume 11, Article number: 608 (2020) https://doi.org/10.1038/s41467-020-14397-8

Abstract

Rewiring cellular sensors to trigger non-natural responses is fundamental for therapeutic cell engineering. Current designs rely on engineered receptors that are limited to single inputs, and often suffer from high leakiness and low fold induction. Here, we present Generalized Engineered Activation Regulators (GEARs) that overcome these limitations by being pathway-specific rather than input-specific. GEARs consist of the MS2 bacteriophage coat protein fused to regulatory or transactivation domains, and work by rerouting activation of the NFAT, NFκB, MAPK or SMAD pathways to dCas9-directed gene expression from genomic loci. This system enables membrane depolarization-induced activation of insulin expression in β-mimetic cells and IL-12 expression in activated Jurkat cells, as well as IL-12 production in response to the immunomodulatory cytokines TGFβ and TNFα in HEK293T cells. Engineered cells with the ability to reinterpret the extracellular milieu have potential for applications in immunotherapy and in the treatment of metabolic diseases.

Page 101 of 200

Introduction

Numerous active clinical trials are investigating the therapeutic potential of engineered cell therapies, mainly in oncology and organ regeneration19. Chimeric antigen receptor T (CAR-T) cells were successfully applied for killing lymphoma cells and have been FDA-approved15. Current research for enhancing CAR-T cell function is mainly focused on improving synthetic receptors and on conditional expression of immunostimulatory cytokines, such as interleukin 2 (IL-2) and interleukin 12 (IL-12), to combat solid tumors more effectively30,191-193. Besides immunotherapy, engineered cells have been applied in many other proof-of-concept treatment strategies in animal disease models, including the treatment of metabolic diseases using encapsulated cells8,46, and therapeutic agent delivery using transgenic mesenchymal stem cells194,195. However, current designs of engineered cells are still limited by the lack of efficient and versatile molecular devices. New methods for precise control of cellular signaling pathways are needed to further improve cell therapies40,193,196-200. Recent efforts have led to the development of synthetic receptors for endogenous promoter activation in response to vascular endothelial growth factor (VEGF), which is important for tumor growth and immunoevasion201,202. However, these systems do not respond to inducer concentrations that typically activate native receptors, and are excluded from natural control mechanisms. Natural signal transduction relies on a multitude of signaling pathways that have evolved for efficient and tightly regulated dynamic behavior. Additionally, many signaling pathways interact with each other, forming multidimensional networks that integrate various inputs to generate sophisticated situation-dependent responses. Therefore, precise quantitative and temporal regulation is critical for obtaining desired effects203. Rewiring natural receptors to specific transgene expression has been established for several proof-of-concept cellular therapies11,46. However, rewiring these pathways directly to endogenous targets remains a challenge. Here, we present a modular system for the direct repurposing of endogenous signaling pathways to activate native or synthetic promoters. This system relies on molecular devices called Generalized Engineered Activation Regulators (GEARs). Each GEAR consists of a MS2 bacteriophage coat protein (MCP) fused to natural regulatory or transactivation domains of key signaling pathways. GEARs combined with catalytically dead CRISPR-associated protein 9 (dCas9) and a synthetic guide RNA (sgRNA) containing two MS2 coat protein-binding loops (MS2)21 (Fig. 1a) can be used to reroute calcium signaling, the TGFβ/SMAD pathway, the NFκB

Page 102 of 200

pathway, and the MAPK/ERK pathway (Fig. 1b). The proof-of-concept applications include membrane depolarization-induced activation of insulin expression in β-mimetic cells, IL-12 expression in activated immortalized human T-lymphocytes (Jurkat), and activation of IL-12 production in response to immunosuppressive TGFβ or immunostimulatory TNFα in HEK293T cells. GEARs have potential for applications in therapeutic cell engineering, especially in immunotherapy and in the treatment of metabolic diseases.

Page 103 of 200

Figure 1 | GEAR toolbox. (a) Catalytically inactive Cas9 (dCas9) binds a DNA sequence complementary to single-stranded guide RNAs equipped with two MS2 aptamers (sgRNA-MS2). GEARs can bind the MS2 aptamers via MS2 coat protein (MCP) domains and activate gene expression in response to endogenous signaling. (b) All GEAR variants contain MCP at the N- terminal end. GEARNFAT contains a transactivation domain (TA) and the regulatory domain of

NFAT (NFATreg). Plasmid pKK50 encodes GEARNFAT under control of human cytomegalovirus promoter (PhCMV-GEARNFAT-pA). GEARSMAD2 contains full-length SMAD2. Plasmid pKK136 encodes GEARSMAD2 under control of human cytomegalovirus promoter (PhCMV-GEARSMAD2-pA).

GEARp65 contains NFκB transcription regulator subunit p65. Plasmid pKK122 encodes GEARp65 under control of human cytomegalovirus promoter (PhCMV-GEARp65-pA). GEARElk contains the transactivation domain of Elk1. Plasmid pKK82 encodes GEARElk under control of human cytomegalovirus promoter (PhCMV-GEARElk-pA). pA is a polyadenylation signal.

Page 104 of 200

Results

GEARNFAT enables rewiring of intracellular calcium signaling Calcium is one of the most pivotal second messengers, intersecting with numerous signaling pathways to mediate multiple cellular processes204. Therefore, we first focused on nuclear factor of activated T-cells (NFAT), which is dephosphorylated upon activation of a calcium-dependent calmodulin-calcineurin cascade, and designed GEARNFAT, which consists of

MCP fused to the transactivation domains of p65 and HSF1 (p65TA-HSF1TA), and the regulatory domain of NFAT (NFATreg). GEARNFAT translocates to the nucleus (Supplementary Fig. 1) and activates sgRNA-specified gene expression in response to an increase of the intracellular calcium level. We used GEARNFAT to upregulate insulin expression upon cell-membrane depolarization in HEK293T cells, mimicking the response observed in pancreatic β-cells205 (Fig. 2a) and evaluated the system performance by means of a reporter gene assay. HEK293T cells were engineered to express L-type voltage gated calcium channel (LVGCC) in order to enhance membrane 46 depolarization-driven calcium influx . Incorporating GEARNFAT into these cells enabled membrane depolarization-dependent expression of a reporter gene controlled by an insulin promoter (Fig. 2b). GEARNFAT activation led to significant reporter gene expression within 8 hours, reaching a maximum after 36 hours (Supplementary Fig. 2). Furthermore, when expressed 46 in a β-mimetic cell line , GEARNFAT increased endogenous insulin gene expression (Fig. 3a, b) and protein production (Fig. 3c) in response to membrane depolarization.

Page 105 of 200

Page 106 of 200

Figure 2 | GEAR-driven transgene expression. HEK293T cells containing a reporter plasmid for human insulin promoter (PhINS) and expressing GEARs, dCas9 and a human insulin promoter specific sgRNA-MS2. At 24 hours after the beginning of transfection, cells were stimulated for 36 hours with the indicated inducer and then the reporter protein SEAP was quantified in the cell culture supernatant. (a, b) Membrane depolarization-dependent GEARNFAT activation. Cells expressing GEARNFAT and the L-type voltage gated calcium channel were depolarized with 40 mM potassium chloride (KCl). (c, d) GEARNFAT activation upon stimulation with vascular endothelial growth factor 165 (VEGF165). Cells expressing GEARNFAT and VEGF receptor 1 (VEGFR1) were stimulated with 20 ng/ml VEGF165. (e,f) Transforming growth factor β (TGFβ)- induced GEARSMAD2 activation. Cells expressing GEARSMAD2 and TGFβ receptor 2 (TBRII) were stimulated with 10 ng/ml TGFβ. (g,h) GEARp65 activation upon stimulation with tumor necrosis factor α (TNFα). Cells transfected with GEARp65 and IκB-encoding plasmids in a 1:1 ratio were stimulated with 10 ng/ml TNFα. (i,j) GEARElk1 activation upon basic fibroblast growth factor (bFGF) stimulation. Cells expressing GEARElk1 were stimulated with 20 ng/ml bFGF. (k,l) GEARElk activation upon chimeric receptor stimulation. Cells expressing GEARElk1 and MAPK-GEMS were stimulated with the synthetic azo-dye RR120 (100 ng/ml). Green bars represent SEAP concentrations measured in the supernatant of stimulated cells (mean values). Blue bars represent controls without inducer. Black dots correspond to individual data points of n = 3 biological replicates. ** p<0.01, *** p<0.001, **** p<0.0001. Statistical significance was calculated using a two-tailed t-test. A detailed description of the statistical analysis is provided in Supplementary Table 5. Source data are provided as a Source Data file.

Page 107 of 200

Figure 3 | Rewiring intracellular calcium signaling to genomic targets. (a) Membrane depolarization-triggered insulin expression. Cell membrane depolarization causes calcium influx via L-type voltage-gated calcium channels, followed by nuclear translocation of GEARNFAT, and endogenous insulin promoter (PhINS) activation. (b, c) 24 hours after the beginning of transfection,

β-mimetic cells expressing dCas9, GEARNFAT and a human insulin promoter-specific sgRNA were depolarized with 40 mM KCl. Insulin mRNA (b) and protein in the supernatant (c) were quantified after 36 hours. (d) GEARNFAT-driven multiple gene activation. β-Mimetic cells expressing dCas9,

L-type voltage-gated calcium channel, GEARNFAT and PhINS-specific sgRNA, as well as PhIL-12B- specific sgRNA were depolarized with 40 mM KCl. Insulin and IL-12B mRNA levels were quantified after 36 h. Blue bars and the left axis correspond to insulin mRNA. Violet bars and the right axis correspond to IL-12B mRNA. Black dots correspond to individual data points of n = 3 biologically independent samples. BDL – below detection limit, ns – nonsignificant, *p<0.05, **p<0.01, ***p<0.001. Statistical significance was calculated using a two-tailed t-test. A detailed description of the statistical analysis is provided in Supplementary Table 5. Source data are provided as a Source Data file.

Page 108 of 200

Next, we tested the multiplexing capability of GEARs and simultaneously transfected dCas9, GEARNFAT and sgRNAs in order to upregulate gene expression from the native genomic loci of insulin and IL-12. The gene expression levels of both insulin and IL-12 were highly upregulated upon cell stimulation with KCl (Fig. 3d). VEGF signaling is another example of an important pathway that is responsive to transient increases of cytosolic calcium concentration. Therefore, we speculated that stimulation of the

VEGF receptor (VEGFR) could also be used to activate GEARNFAT. We found that GEARNFAT was functional in VEGFR-expressing HEK293T cells and led to robust transgene expression (Fig. 2c, d), as well as endogenous gene expression (Supplementary Fig. 3a) in response to stimulation with VEGF. Similarly, rewiring of VEGFR using GEARNFAT was functional in human mesenchymal stem cells (hMSCs-TERT) (Supplementary Fig. 4a). This response uses the naturally optimized behavior of native receptors and endogenous signaling to achieve the high sensitivity that is observed in nature. Direct comparison of the VEGFR/GEARNFAT system with 202 previously published synthetic receptors (MESA) showed the superiority of GEARNFAT under the tested conditions (Supplementary Fig. 5). NFAT was originally discovered as a mediator of the responses of activated T-cells206. In the native pathway, stimulation of the T-cell receptor complex leads to increased cytosolic calcium concentration, followed by the activation of T-cell responses. Thus, expressing GEARNFAT allowed us to hijack calcium signaling in a T-cell line (Jurkat) to increase the expression of endogenous immunostimulatory cytokine IL-12207 (Fig. 4a), thereby illustrating the potential value of GEARs in the context of immunotherapy. We confirmed that GEARNFAT mediated upregulation of IL-12 at both the mRNA and the protein level in response to increased cytosolic calcium concentration (Fig. 4b, c).

Page 109 of 200

Figure 4 | Interleukin 12 expression upon T-cell activation. (a) Schematic representation of T- cell activation. T-cell receptor (TCR) signaling is transduced via calcium. An increase in the intracellular calcium level activates GEARNFAT and leads to elevated expression of interleukin 12

(IL-12). (b, c) Jurkat cells stably expressing dCas9, GEARNFAT and a human IL-12B promoter- specific sgRNA, or an insulin promoter-specific sgRNA as the negative control, were activated with 2 µg/mL ionomycin and 20 ng/mL PMA. (b) IL-12B mRNA was quantified after 8 h. (c) IL- 12B protein was quantified from the cell culture supernatant after 2, 4 and 8 hours. Black dots correspond to individual data points of n = 3 biological replicates. ns – nonsignificant, **p<0.01, ****p<0.0001. Statistical significance was calculated using a two-tailed t-test. A detailed description of the statistical analysis is provided in Supplementary Table 5. Source data are provided as a Source Data file.

Page 110 of 200

We next examined whether GEARNFAT influences the expression levels of native NFAT- dependent genes by means of a reporter gene assay, as well as by qPCR analysis of known NFAT- responsive genes. We found that GEARNFAT downregulated expression of NFAT-driven genes in a GEARNFAT-positive stable cell line of Jurkat cells (Supplementary Fig. 6), as well as in

HEK293T cells transfected with a PNFAT-driven reporter plasmid (Supplementary Fig. 7). The downregulation of NFAT-driven genes was dependent on the amount of the transfected

GEARNFAT-encoding plasmid: increasing the amount of the plasmid increases production of the

GEAR-controlled reporter protein (SEAP), while decreasing the production of PNFAT-driven NanoLuc luciferase (Supplementary Fig. 7). In the design of GEARs, high flexibility and modularity are achieved by the physical separation of a regulatory module and the dCas9-dependent DNA-binding module. For the regulatory module, we took advantage of the small size of the MCP protein, because we speculated that a smaller fusion protein might be more efficiently regulated by the cellular machinery, e.g. for induced nuclear translocation. In a direct comparison of GEARNFAT and a construct composed of dCas9 fused to the regulatory domain of NFAT and a transactivator (named CaRROT)207,

GEARNFAT outperformed CaRROT under the tested conditions (Supplementary Fig. 2a, Supplementary Fig. 8).

Generalizing functionality of GEARs to other signaling pathways

We next generalized the functional principle of GEARNFAT to incorporate other signaling pathways. We engineered GEARSMAD2 to transduce signals from receptors of the transforming growth factor beta (TGFβ) superfamily. The functionality of the construct was confirmed with

PhINS-driven reporter gene expression in response to stimulation with TGFβ in HEK293T cells (Fig. 2e, f) and hMSCT-TERT cells (Supplementary Fig. 4b) transiently expressing the TGFβ receptor, as well as with activation of the endogenous insulin promoter (Supplementary Fig. 3b). As proof of concept for converting the normally immunosuppressive effect of TGFβ into an immunostimulatory response, TGFβ signaling was rewired to the expression of IL-12 (Fig. 5a, b).

Page 111 of 200

Figure 5 | Rewiring inflammation-related signaling pathways to genomic targets. (a) Immunomodulatory signal conversion. TGFβ-induced TBRII activation causes nuclear translocation of GEARSMAD2 and subsequent IL-12 promoter (PhIL-12) activation. (b) At 24 hours after the beginning of transfection, HEK293T cells expressing dCas9, GEARSMAD2, TBRII and a human IL-12B promoter-specific sgRNA were stimulated with 10 ng/ml TGFβ. IL-12B mRNA was quantified after 36 h. (c) Immunomodulatory signal switch. TNFα-induced TNFR activation causes IκB degradation, resulting in nuclear translocation of GEARp65 and IL-12 promoter (PhIL-

12) activation. (d) At 24 hours after the beginning of transfection, HEK293T cells expressing dCas9, GEARp65, IκB and a human IL-12B promoter specific sgRNA were stimulated with 10 ng/ml TNFα. IL-12B mRNA was quantified after 36 hours. Violet bars represent mRNA expression (relative to GAPDH) in stimulated cells (mean value). Blue bars represent mRNA expression in controls without inducer. Black dots correspond to individual data points of n = 3 biological replicates. * p<0.05, ** p<0.01. Statistical significance was calculated using a two- tailed t-test. A detailed description of the statistical analysis is provided in Supplementary Table 5. Source data are provided as a Source Data file.

Page 112 of 200

Next, we generated the NFκB-dependent GEAR (GEARp65) by fusing MCP to the NFκB subunit p65. NFκB nuclear translocation depends on the degradation of its binding partner IκB, which exposes the nuclear localization signal (NLS) of p65. GEARp65 was therefore coexpressed with IκB to avoid non-induced p65 activity. We used GEARp65 to redirect the signal of TNFα. Stimulation with TNFα leads to the activation of endogenously expressed receptors and increased

PhINS-driven transgene activation in HEK293T (Fig. 2g, h) and hMSC-TERT (Supplementary

Fig. 4c). GEARp65 targeted for PhINS (Supplementary Fig. 3c) or PhIL12 increased endogenous insulin or interleukin 12 (IL-12) expression, respectively (Fig. 5c, d). This effect might be relevant for immunotherapy, as IL-12 is an important immunostimulatory cytokine, but its expression is inhibited upon exposure to TNFα in some cell types208,209.

MEK/ERK-activated GEAR (GEARElk1) consists of MCP fused to the Elk1 transactivation domain (Elk1TA). GEARElk (PhCMV-GEARElk1-pA, pKK82) was used to rewire signaling of fibroblast growth factor (FGF), an important signaling molecule overexpressed in certain tumors210. Incubation with basic FGF (bFGF) leads to increased transgene expression in HEK293T (Fig 2i, j) and hMSC-TERT (Supplementary Fig. 4d). Moreover, the system is compatible with the recently developed Generalized Extracellular Molecule Sensor (GEMS)211 platform (Fig. 2k, l). However, GEARElk1 did not increase expression of tested endogenous genes: insulin, IL-12 or interleukin 2 (IL-2).

Functional characterization and non-specific effects Control experiments conducted without GEARs or with unspecific sgRNAs showed little or no increase in reporter gene expression, or in endogenous gene activation that could be attributed to the native host transcription machinery, while expression of GEARs together with specific sgRNAs also increased non-induced reporter gene expression (Supplementary Fig. 9, Supplementary Fig. 10). We further evaluated the effect on the expression of endogenous genes, focusing on insulin as a gene typically not expressed in HEK293T cells, as well as TGFβ, which is highly expressed in this cell type. The insulin transcript was not detected in the absence of GEARs. The transcription level of TGFβ did not show any major change due to binding the non- activating dCas9 complex (Supplementary Fig. 11). Dose-response experiments performed for

VEGF, TGFβ, TNFα and bFGF (using systems incorporating GEARNFAT, GEARSMAD2, GEARp65 and GEARElk1 respectively) confirmed that the sensitivity of GEAR-based systems depends on the sensitivity of the receptor (Supplementary Fig. 12). Their kinetic profiles showed that activation

Page 113 of 200

was observable after 4 hours (Supplementary Fig. 2). Since reporter gene assay results (Fig. 2) do not always correlate well to RNA content quantified by qPCR (Fig. 3), we performed both assays. These assays showed qualitatively similar increases in SEAP reporter gene activation at the mRNA and protein levels (Supplementary Fig. 13).

Discussion

Therapeutic cell engineering requires various molecular devices and methods to achieve the desired cell behavior. Many approaches involve regulating transcription by employing orthogonal212,213 or native44,46,214 signaling pathways, or combinations of the two46,56. Native cell signaling pathways are optimized by evolution to precisely regulate cellular behavior. Hijacking a signaling node that is shared by multiple pathways thus enables us to add a novel function to an operational system, such as expression of IL-12 concomitantly with T-cell line activation. Hence, GEARs both complement and extend the common trend in synthetic biology to design inducer- specific systems215. Development of dCas9 greatly facilitated the conditional modulation of endogenous gene expression. Traditional approaches, which involve inducible expression of the dCas9 protein, typically require more than 2 days from the activation time until the appearance of a detectable effect, and the effect lasts for hours to days216-219. The dynamics is slowed down by the need for a two-step transcriptional control system, consisting of transcription of dCas9 followed by transcription of target genes, as well as the stability (in the timescale of hours) of dCas9 mRNA, which slows the off-kinetics. GEAR components can bypass both delays by eliminating the requirement of inducible dCas9 expression, and by following the dynamics of the endogenous signaling cascade. The therapeutic effect of engineered cells often depends on the function of the whole cell, rather than on the expression of a single gene. Jurkat cells with stable incorporation of GEARNFAT maintain the expression of native NFAT-dependent genes related to cell activation, as shown by the upregulation of IL-2 and other tested genes. In general, gene expression levels are controlled by numerous factors, which include the promoter sequence, the epigenetic state of chromatin, and the activity of related signaling pathways. Typically, linking a signaling pathway to transgene expression increases its basal level, even in the non-activated state31,33,34. The magnitude of that effect depends on the signaling pathway, as it was observed for GEAR-driven reporter gene expression (time-dependent increase visible in Supplementary Fig. 2), as well as on the

Page 114 of 200

expression strength of the GEAR (Supplementary Fig. 7a). Simultaneously, GEARs sequester parts of the transcription machinery, leading to decreased expression of genes controlled by natural transcription factors (Supplementary Fig. 7b, Supplementary Fig. 6). dCas9-based activation of genomic targets that are additionally regulated by epigenetics is typically context-dependent and can be affected by the targeted region, as well as by the choice of transactivators188,220. The dCas9 complex might interfere with natural transcription factor-DNA interactions by steric hindrance, leading to either inhibition or activation of transcription221-223, which opens further possibilities to tune the system. As with other DNA-targeting gene switches, there could be possible off-target effects224 that could be minimized by the careful choice of the targeted region, the signaling pathway and the GEAR expression level. Rerouting or tapping into endogenous signaling pathways may provide engineering opportunities to create designer cells and adapt their behavior for therapeutic purposes. Synthetic receptors are another type of molecular devices that are currently used in therapeutic cell engineering. They can sense a soluble input and transduce the signal either by viral protease-mediated release of an effector protein, or by activating native signaling pathways. Typically, TEV protease-based systems require high inducer concentrations to trigger receptor dimerization and thus may not interface optimally with natural control mechanisms. GEARs, on the other hand, respond to inducer concentrations in the same range as native receptors. In a direct comparison of GEARNFAT-driven gene expression in response to VEGF receptor activation with 202 the TEV protease-based synthetic MESA receptor system GEARNFAT showed superior performance in terms of achieved fold induction and sensitivity. The second class of synthetic receptors, which includes GEMS211, relies on native signaling pathways. Hence, GEARs can be combined with GEMS to form artificial signaling pathways for normally non-signaling molecules. We believe GEARs have great potential to extend the functionality of therapeutic cells, and could also find a variety of applications in basic research, especially in studies of cell signaling and systems biology.

Page 115 of 200

Acknowledgements. We thank Viktor Hällman for sharing figure templates, Gina Melchner von Dydiowa for sharing plasmids, Pratik Saxena for his advice and helpful discussions, Erik Aznauryan for sharing Jurkat cells.

Author contributions. K.K. and L.S. designed the project. K.K., L.S. and M.F. wrote the manuscript, K.K., L.S., H.K. designed and performed the experiments. K.K., L.S., and M.F. analyzed the results.

Competing interests statement. The authors declare no competing interests.

Page 116 of 200

Methods

Genetic constructs used in this study. Comprehensive design, construction details for all original expression vectors are provided in Supplementary Table 1. Targeting sequences of synthetic guide RNA are listed in Supplementary Table 2.

Guide RNA design. sgRNA target sequences were designed using Benchling (Benchling [Biology Software]. (2019). Retrieved from [https://benchling.com]). This software calculates on-target and off-target scores based on algorithms developed by Doench et al.225 and Hsu et al.226.

Cell culture and transfection. Human embryonic kidney cells (HEK293T, ATCC: CRL-11268) were obtained from ATCC and cultivated in Dulbecco’s modified Eagle’s medium (DMEM; cat. no. 52100-039; Thermo Fischer Scientific, Waltham, MA, USA) supplemented with 10 % fetal bovine serum (FBS; cat. no. F7524, lot no. 022M3395, Sigma-Aldrich), 100 U/mL penicillin and 100 µg/mL streptomycin (penicillin-streptomycin solution 100x; cat. no. L0022, Biowest, Nuaillé,

France) and grown at 37C in a humidified atmosphere containing 5 % CO2. For transfection, 35’000 cells were seeded per cm2 of the cell culture dish and, after 24 h, incubated for another 6 h with a 1:6 DNA:PEI (Polyethylenimine MAX; MW 40,000, cat. no. 24765-2; Polysciences Inc., Warrington, PA, USA) solution containing 1.5 µg DNA per cm2 of transfected cells. Immortalized human mesenchymal stem cells (hMSC-TERT) were a kind gift from Moustapha Kassem (University Hospital of Aarhus and University Hospital of Odense, Denmark). They were cultivated and transfected the same way as HEK-293T. Jurkat cells (Jurkat, Clone E61, ATCC: TIB152™) were a kind gift from Sai Reddy (ETHZ, Switzerland). Jurkat cells were cultivated in RPMI medium (Cat. no. 72400-021; Thermo Fischer Scientific, Waltham, MA, USA) supplemented with 10 % fetal bovine serum 100 U/mL penicillin and 100 µg/mL streptomycin. Transfection was performed using Xfect Transfection Reagent (Cat. no. 631318; Takara Bio Europe SAS, St Germain en Laye, France) according the manufacturer’s instructions. Briefly, 2 million cells were resuspended in 2 ml of fresh media and seeded into a 35 mm dish. 5 μg of plasmid DNA was mixed with 1.5 µl of Xfect reagent, incubated for 10 minutes and added to the cells. Detailed transfection protocols are provided in Supplementary Table 3. None of the cell lines used in this study is listed in the Register of Misidentified Cell Lines, ICLAC. Cell lines were authenticated by means of microscopy and by activation-dependent IL-2 expression for Jurkat cells. Cells were not tested for mycoplasma contamination.

Page 117 of 200

Stable cell lines. Jurkat cells were transfected with plasmids encoding SB100X expression vector pCMV-T7-SB100 (PhCMV-SB100X-pA) and the SB100X-specific transposon-containing plasmids pKK171 (ITR-PhCMV-dCas9-pA:PRPBSA-dTomato-P2A-BlastR-pA-ITR) and pKK172 (ITR-

PhCMV- GEARNFAT -pA:PRPBSA-dTomato-P2A-BlastR-pA-ITR). After 48 hours, cells were selected with conditioned medium containing 5 µg/mL blasticidin for 21 days. Conditioned medium was prepared by sterile filtration of cell culture medium from wild-type Jurkat cells after 48 hours in culture. Next, blasticidin-resistant cells were transfected with pCMV-T7-SB100 and pLeo1209

(ITR-PhU6-gRNAIL-12-pA-ITR:PRPBSA-BFP-P2A-PuroR-pA-ITR) and 48 hours later, selected with conditioned medium containing 0.5 µg/mL puromycin for 21 days. Stably transgenic puromycin- resistant cell populations were selected by FACS-mediated single-cell sorting based on cellular fluorescence relative to wild-type Jurkat (BFP: 405 nm laser, 450/50 bandpass filter; dTomato: 561 nm laser, 570 nm long-pass filter, 586/15 bandpass filter) using a Becton Dickinson LSRII Fortessa flow cytometer (Becton Dickinson, Allschwil, Switzerland).

Cell stimulation. Transfected HEK293T cells were washed with DMEM containing 1% FBS, 100 U/mL penicillin and 100 µg/mL streptomycin for 18 h, and incubated with the concentrations indicated in the respective figures for 36 h (unless stated otherwise) with one of the following reagents: recombinant human VEGF165 (Cat. no. 100-20, Peprotech, London, United Kingdom), recombinant human FGF-basic (154 a.a.) (Cat. no. 100-18B, Peprotech, London, United Kingdom), recombinant human TNFα (Cat. no. 300-01A, Peprotech, London, United Kingdom), or recombinant human TGF-β1 (Cat. no. 100-21, Peprotech, London, United Kingdom). Transfected hMSC-TERT cells were stimulated in DMEM containing 10% FBS, 100 U/mL penicillin and 100 µg/mL streptomycin. Jurkat cells were centrifuged and resuspended in fresh growth medium containing the indicated concentrations of ionomycin (Cat. no. I0634-1MG, Sigma-Aldrich) and phorbol 12-myristate 13-acetate (PMA; Cat. no. P1585-1MG, Sigma- Aldrich).

Analytical assays. SEAP (human placental secreted alkaline phosphatase) levels were profiled in cell culture supernatants using a colorimetric assay. 100 µL 2x SEAP assay buffer (20 mM homoarginine, 1 mM MgCl2, 21% diethanolamine, pH 9.8) was mixed with 80 µL heat-inactivated (30 min at 65°C) cell culture supernatant. After the addition of 20 µL substrate solution (120 mM p-nitrophenyl phosphate; cat. no. AC128860100, Thermo Fisher Scientific, Waltham, MA, USA), the absorbance time course was recorded at 405 nm and 37°C using a Tecan Genios PRO plate reader (cat. no. P97084; Tecan Group AG, Maennedorf, Switzerland) and the SEAP levels

Page 118 of 200

were determined as follows: first, absorbance change over time (slope) was calculated. According to the Beer-Lambert’s law, absorbance is proportional to the concentration of a colored compound and depends on the light path length (d) and extinction coefficient (ε) (ε for p-nitrophenyl (εpNP) = 18.600 M-1cm-1). Enzymatic activity EA [U/L] was calculated from the equation: EA = slope * -1 -1 227 dilution factor * εpNP * d .

NanoLuc® luciferase was quantified in cell culture supernatants using the Nano-Glo® Luciferase Assay System (cat. no. N1110; Promega, Duebendorf, Switzerland). In brief, 7.5 µL of cell culture supernatant was added per well of a black 384-well plate and mixed with 7.5 µL substrate- containing assay buffer. Total luminescence was quantified using a Tecan Genios PRO plate reader (Tecan Group AG).

ELISA. Human Insulin was quantified with Mercodia Insulin ELISA (Cat. no. 10-1113-01, Mercodia, Uppsala, Sweden). Interleukin 12 (IL-12) was quantified using human IL-12 (p40) ELISA Kit (Cat. no. KAC1561, Thermo Fisher Scientific, Waltham, MA, USA).

Quantitative RT–PCR. Total RNA of HEK293T cells was isolated using the Quick-RNA kit (Zymo Research, Irvine, CA, U.S.A.). Reverse transcription was performed using a High-Capacity cDNA Reverse Transcription Kit (Cat. no. 4368814, Thermo Fisher Scientific, Waltham, MA, USA). Quantitative PCR was performed using SsoAdvanced Universal SYBR® Green Supermix (Cat. no. 1725270, Bio-Rad, Hercules, CA, USA). The Eppendorf Realplex Mastercycler (Eppendorf GmbH) was set to the following amplification parameters: 30 s at 95 °C and 40 cycles of 15 s at 95 °C followed by 30 s at X °C (X = 59 for insulin, 64 for IL-12, 59 or 64 for GAPDH).

The relative threshold cycle (Ct) was determined and normalized to the endogenous glyceraldehyde 3-phosphate dehydrogenase (GAPDH) transcript. The fold change for each transcript relative to the control was calculated using the comparative Ct method. qPCR primer pairs are listed in Supplementary Table 4.

Cell imaging. Cells were fixed using 2% paraformaldehyde in phosphate-buffered saline (PBS) for 20 minutes. Microscope images were taken using a Nikon Ti-U wide-field inverted microscope equipped with a Nikon Intensilight. Green fluorescence was recorded using a 469/35 nm excitation filter, 495 nm long pass filter, and 520/35 nm emission filter. Images were processed using the ImageJ package.

Data representation and statistics. Representative data are presented for each figure as individual values (black dots) and mean values (bars). n = 3 refers to biological replicates.

Page 119 of 200

Statistical significance was calculated with a two-tailed t-test, using GraphPad Prism 7.04 software. Normal distributions were assumed. Variance equality was tested using the f-test and Welch’s correction to the t-test was applied when the variances of compared data sets were significantly different (p<0.05). Whenever the result of an assay was below the detection limit (BDL), the value “0” was assigned to that sample to calculate statistical significance. The source data is provided as a Source Data file.

Data availability. The authors declare that all the data supporting the findings of this study are available within the paper and its supplementary information files. Sequence data of original plasmids have been deposited in GenBank under accession codes MN811100 [https://www.ncbi.nlm.nih.gov/nuccore/MN811100] - MN811119 [https://www.ncbi.nlm.nih.gov/nuccore/MN811119] and original plasmids are available upon request. All vector information is provided in Supplementary Table 1. Numerical data and detailed statistical analysis for each figure is provided in Supplementary Table 5. Source data is provided in the source data file.

Page 120 of 200

Supplementary Information

Supplementary Figure 1. Nuclear translocation GEARNFAT

Supplementary Figure 2. Kinetics of GEARNFAT -mediated transgene expression

Supplementary Figure 3. Endogenous insulin expression

Supplementary Figure 4. GEAR-mediated transgene expression in immortalized human mesenchymal stem cells

Supplementary Figure 5. Synthetic receptors MESA

Supplementary Figure 6. Unspecific effect of GEARNFAT expression on endogenous gene transcription in Jurkat cells.

Supplementary Figure 7. GEARNFAT plasmid titration

Supplementary Figure 8. Calcium-inducible dCas9-based nuclear translocation protein “CaRROT”

Supplementary Figure 9. Negative control experiments – transgene expression

Supplementary Figure 10. Negative control experiments – endogenous gene expression

Supplementary Figure 11. Effect of GEAR expression on the basal transcription of endogenous genes

Supplementary Figure 12. Dose-response relationships

Supplementary Figure 13. SEAP assay and qPCR comparability

Supplementary Table 1. Plasmids used and designed in this study

Supplementary Table 2. Synthetic guide RNA sequences

Supplementary Table 3. Detailed transfection protocols

Supplementary Table 4. qPCR primer pairs

Supplementary Table 5. Detailed statistics

Page 121 of 200

Page 122 of 200

Supplementary Figure 1. Nuclear translocation of GEARNFAT. β-Mimetic cells (CaV1.3- transgenic HEK293T) transfected with plasmids encoding GEARNFAT were depolarized with 40 mM KCl for 4 hours. Images were captured with a wide-field fluorescence microscope. Panels a, b, c and g show non-induced cells (Negative Control). Panels d, e, f, and h show depolarized cells (40 mM KCl). Green fluorescence of GFP-tagged GEARNFAT is presented in panels a, d, g, and h, while panels b and e show blue fluorescence of DAPI-stained nuclei. Panels c and f present merged channels for non-induced and depolarized cells, respectively. Panels a-f were captured using a 10x magnification objective. For panels g and h, a 40x magnification objective was used.

Page 123 of 200

Page 124 of 200

Supplementary Figure 2. Kinetics of GEAR-mediated transgene expression. Cells were transfected with plasmids encoding human insulin promoter-specific sgRNA (sgRNAINS), SEAP reporter plasmid controlled by human insulin promoter (PhINS-SEAP), dCas9, an indicated GEAR and an indicated receptor. At 48 hours post-transfection, cells were stimulated with a compound indicated in the panel and SEAP was quantified from the supernatant after 4, 8, 24 and 36 hours.

(a) Membrane depolarization (40 mM KCl) of β-mimetic cells (CaV1.3-transgenic HEK293T) and

GEARNFAT activation. (b) 10 ng/ml VEGF165-mediated GEARNFAT activation. (c) 10 ng/ml

TGFβ-mediated GEARSMAD2 activation. (d) 10 ng/ml TNFα-mediated GEARp65 activation. (e)

10 ng/ml bFGF-mediated GEARElk1 activation. Numbers above the data points indicate fold induction in the stimulated cells (green line) compared to the negative control (blue line). Black dots correspond to mean. Error bars show the standard error of the mean (SEM). n = 3 biological replicates. Source data are provided as a Source Data file.

Page 125 of 200

Supplementary Figure 3. Endogenous insulin expression. HEK293T cells expressing dCas9 and PhINS-specific sgRNA (sgRNAINS) were stimulated for 36 h with the indicated inducers. Insulin mRNA levels were quantified in relation to GAPDH. (a) VEGF165-induced GEARNFAT. (b)

TGFβ-induced GEARSMAD2. (c) TNFα-induced GEARp65. Violet bars represent mRNA expression (relative to GAPDH) in stimulated cells (mean value). Blue bars represent mRNA expression in controls without inducer. Black dots correspond to individual data points of n = 3 biological replicates. BDL – below detection limit. * p<0.05, **** p<0.0001. Statistical significance was calculated using a two-tailed t-test. A detailed description of the statistical analysis is provided in Supplementary Table 5. Source data are provided as a Source Data file.

Page 126 of 200

Supplementary Figure 4. GEAR-mediated transgene expression in immortalized human mesenchymal stem cells. hMSC-TERT cells containing a SEAP reporter plasmid under control of the human insulin promoter (PhINS-SEAP) and expressing dCas9, as well as the indicated receptors and GEARs were stimulated for 36 hours with the indicated inducers. SEAP was quantified from the cell culture supernatant. (a) VEGF receptor 1 (VEGFR1) and GEARNFAT, (b)

TNFα receptor (TNFR), GEARp65 and IκB, (c) TGFβ receptor II (TBRII) and GEARSMAD2, (d) endogenous FGF receptor (FGFR) and GEARElk1. Black dots correspond to individual data points of n = 3 biological replicates. *p<0.05, ***p<0.001. Statistical significance was calculated using a two-tailed t-test. A detailed description of the statistical analysis is provided in Supplementary Table 5. Source data are provided as a Source Data file.

Page 127 of 200

Page 128 of 200

Supplementary Figure 5. Synthetic receptors, MESA. HEK293T cells were transfected with plasmids encoding a human insulin promoter-specific sgRNA (sgRNAINS), and (a,b) a SEAP reporter plasmid controlled by tetracycline repressor binding sites (OTetR-PhCMVmin-SEAP), plasmids for V1 MESA tTA and V1 MESA TEV, (c,d) a SEAP reporter plasmid controlled by the human insulin promoter (PhINS-SEAP), and plasmids for V1 MESA dCas9-VP64 and V1 MESA

TEV, (e,f) a SEAP reporter plasmid controlled by tetracycline repressor binding sites (OTetR-

PhCMVmin-SEAP), and plasmids for V2 MESA tTA and V2 MESA TEV, (g,h) a SEAP reporter plasmid controlled by human insulin promoter (PhINS-SEAP), and plasmids for V2 MESA dCas9- VP64 and V2 MESA TEV. Cells were induced with the indicated concentrations of recombinant mouse VEGF-164 (mVEGF). At 36 hours after stimulation, SEAP was quantified in the supernatant. Values above the graphs indicate the total amount of two MESA-encoding plasmids per 1 well of a 24-well plate. These plasmids were transfected at a ratio of MESA-transactivator (tTA or dCas9) to MESA-TEV of 24:1, as described by Schwartz et al.1 Black dots correspond to individual data points of n = 3 biological replicates. ns – nonsignificant, *p<0.05. Statistical significance was calculated using a two-tailed t-test. A detailed description of the statistical analysis is provided in Supplementary Table 5. Source data are provided as a Source Data file.

Page 129 of 200

Page 130 of 200

Supplementary Figure 6. Unspecific effect of GEARNFAT expression on endogenous gene transcription in Jurkat cells. (a) Wild-type Jurkat cells, or (b) Jurkat cells stably expressing dCas9, GEARNFAT, and interleukin 12B promoter-specific sgRNA (sgRNAIL-12) were stimulated for 8 hours with 0.5 µg/mL ionomycin and 5 ng/mL phorbol 12-myristate 13-acetate (PMA). mRNA levels were quantified relative to GAPDH. Bars represent mean values for stimulated cells (violet bars) and control with no inducer (blue bars). Black dots correspond to individual data points of n = 3 biological replicates. Source data are provided as a Source Data file.

Page 131 of 200

Supplementary Figure 7. GEARNFAT plasmid titration. β-Mimetic cells (CaV1.3-transgenic HEK293T) were transfected with plasmids encoding dCas9 and the indicated amounts of

GEARNFAT, human insulin promoter-specific sgRNA (sgRNAINS), as well as the SEAP reporter plasmid controlled by human insulin promoter (PhINS-SEAP), and the NanoLuc reporter plasmid controlled by an NFAT-dependent promoter (PNFAT-NanoLuc). Cells were depolarized with 40 mM KCl. (a) SEAP and (b) NanoLuc were quantified from the cell culture supernatant after 48 hours. Black dots correspond to individual data points of n = 3 biological replicates. Source data are provided as a Source Data file.

Page 132 of 200

Supplementary Figure 8. Calcium-inducible dCas9-based nuclear translocation protein

“CaRROT”. (a) β-Mimetic cells (CaV1.3-transgenic HEK293T) were transfected with plasmids encoding a human insulin promoter-specific sgRNA (sgRNAINS), as well as the SEAP reporter plasmid controlled by the human insulin promoter (PhINS-SEAP) and CaRROT. At 48 hours post- transfection cells were depolarized with 40 mM KCl and SEAP was quantified from supernatant samples after 4, 8, 24 and 36 hours. Black dots correspond to mean. Error bars show standard error of the mean (SEM). n = 3 biological replicates. Source data are provided as a Source Data file.

Page 133 of 200

Page 134 of 200

Supplementary Figure 9. Negative control experiments – transgene expression. At 24 hours after the beginning of transfection, HEK293T cells containing a SEAP reporter plasmid for human insulin promoter (PhINS-SEAP) and expressing dCas9 were stimulated for 36 hours with the indicated inducers in the presence or absence of GEARs, as well as in the presence or absence of

PhINS-specific sgRNA (sgRNAINS). SEAP was quantified from cell culture supernatant. In No-GEAR control cells, GEARS were replaced by an equal amount of a plasmid expressing MCP- YPet under control of the identical promoter (YPet is a yellow fluorescent protein). sgRNA- negative control cells were transfected with a plasmid encoding a sgRNA targeting human interleukin 2 promoter (sgRNAIL-2). Green bars represent SEAP concentrations measured in the supernatant of stimulated cells (mean values). Blue bars represent controls without inducer. Black dots correspond to individual data points of n = 3 biologically independent samples. (a) Membrane depolarization-induced GEARNFAT. (b) VEGF165-induced GEARNFAT. (c) TGFβ-induced

GEARSMAD2. (d) TNFα-induced GEARp65. (e) bFGF-induced GEARElk1. ns – nonsignificant, *p<0.05, ** p<0.01, *** p<0.001, **** p<0.0001. Statistical significance was calculated using a two-tailed t-test. A detailed description of the statistical analysis is provided in Supplementary Table 5. Source data are provided as a Source Data file.

Page 135 of 200

Page 136 of 200

Supplementary Figure 10. Negative control experiments – endogenous gene expression. At 24 hours after the beginning of transfection, HEK293T cells expressing dCas9 were stimulated for 36 hours with the indicated inducers in the presence or absence of GEARs, as well as in the presence or absence of specific sgRNAs. mRNA levels were quantified in relation to GAPDH. In No-GEAR control cells, GEARS were replaced by an equal amount of a plasmid expressing MCP- YPet under control of the identical promoter. Blue bars represent relative gene expression for controls without inducer (mean values). Violet bars represent relative gene expression for stimulated cells (mean values). Black dots correspond to individual data points of n = 3 biologically independent samples. (a) Endogenous insulin expression. (b) Endogenous interleukin 12 (IL-12) expression. BDL – below detection limit. ns – nonsignificant. *p<0.05, ** p<0.01, *** p<0.001. Statistical significance was calculated using a two-tailed t-test. A detailed description of the statistical analysis is provided in Supplementary Table 5. Source data are provided as a Source Data file.

Page 137 of 200

Supplementary Figure 11. Effect of GEAR expression on basal transcription of endogenous genes. HEK293T cells transfected with plasmids encoding dCas9, the indicated sgRNAs, the indicated GEARs, or not transfected (WT). RNA was isolated and (a) insulin mRNA, or (b) TGFβ mRNA was quantified. Plasmid encoding a fusion protein MCP-YPet (PhCMV-MCP-YPet) was used for the “no GEAR” controls. BDL – below detection limit. Black dots correspond to individual data points of n = 3 biological replicates. Source data are provided as a Source Data file.

Page 138 of 200

Supplementary Figure 12. Dose-response relationships. HEK293T cells containing a reporter plasmid for human insulin promoter (PhINS-SEAP) and expressing dCas9, human insulin promoter specific sgRNA (sgRNAINS), as well as the indicated receptors and GEARs: (a) VEGF receptor 1

(VEGFR1) and GEARNFAT, (b) TGFβ receptor type II (TBRII) and GEARSMAD2 (c) TNFα receptor (TNFR), GEARp65 and IκB, (d) endogenous FGF receptor (FGFR) and GEARElk1. Cells were stimulated for 24 hours with the indicated concentration of the inducer, and the reporter protein SEAP was quantified from the cell culture supernatant. Black dots represent mean. Error bars show standard error of the mean (SEM). n = 3 biological replicates. Solid lines represent four- parameter nonlinear fitted dose-response curves. EC50 – half maximal effective concentration ± 95% confidence range. Source data are provided as a Source Data file.

Page 139 of 200

Supplementary Figure 13. SEAP assay and qPCR comparability. β-Mimetic cells expressing dCas9, GEARNFAT, a human insulin promoter-specific sgRNA (sgRNAINS), as well as a SEAP reporter plasmid controlled by the human insulin promoter (PhINS-SEAP) were depolarized with 40 mM KCl. 24 hours later (a) SEAP mRNA was quantified in the cell lysate and (b) SEAP protein was quantified in supernatant samples. 8.5  and 6.7  correspond to the fold induction values in stimulated cells (40 mM KCl) compared with noninduced cells. Black dots correspond to individual data points of n = 3 biological replicates. * p<0.05. Statistical significance was calculated using a two-tailed t-test. A detailed description of the statistical analysis is provided in Supplementary Table 5. Source data are provided as a Source Data file.

Page 140 of 200

Supplementary Table 1. Plasmids used and designed in this study Plasmid Description and cloning strategy Reference

Constitutive, PhCMV-driven vector expressing calcium-inducible dCas9-based nuclear translocation protein pCaRROT-V5 228 “CaRROT” (PhCMV-CaRROT-V5-pA).

Constitutive α1D subunit of the murine L-type voltage-gated calcium channel CaV1.3 expression vector pCaV1.3 181 (PhCMV-α1D-pA) (Addgene no. 26576).

Constitutive β3 subunit of the murine L-type voltage-gated calcium channel expression vector Lipscombe Lab pCaVb3 (PhCMV-β3-pA) (Addgene no. 26574). (unpublished)

Constitutive α2/δ1 subunits the murine L-type voltage-gated calcium channel expression vector pCaVα2δ1 182 (PhCMV-α2/δ1-pA) (Addgene no. 26575).

Thermo Fisher pcDNA3.1(+) Constitutive PhCMV-driven mammalian expression vector (PhCMV-MCS-pA) Scientific, CA

229 pCEP4YPet-MAMM Constitutive YPet expression vector (PhCMV-YPet-pA) (Addgene no. 14032).

230 pCGN-ELK-1 Constitutive human Elk-1 expression vector (PhCMV-Elk1-pA) (Addgene no. 27156).

189 pCMV-T7-SB100 Constitutive SB100X expression vector (PhCMV-SB100X-pA). (Addgene no. 34879).

Page 141 of 200

Plasmid Description and cloning strategy Reference

Merck, pCOLADuet-1 Prokaryotic expression vector used as a filler plasmid for cotransfection of mammalian cells. Germany

Catalytically dead, human codon-optimized Cas9 under the control of Murine Stem Cell retroVirus LTR pdCas9-humanized 231 promoter (Addgene no. 44246).

Constitutive YPet expression vector (PhCMV-YPet-pA). pFOX8 Unpublished

Constitutive Streptococcus pyogenes catalytically dead CRISPR associated protein 9 (dCas9) expression

vector (PhCMV-dCas9 -pA). dCas9 was PCR-amplified from pdCas9-humanized using OMM223 (5‘- pGM47 gcgaattcaccatgactagtGACAAGAAGTATTCTATCGGAC-3‘) and OMM224 (5‘- This work aagctttctagacaccggtggatccgctagcAGCTCCCTCATCCCCTCCGAGCTG-3‘), digested with EcoRI/BamHI, and cloned into corresponding sites of pMM1 (GenBank accession code MN811115).

PhU6-driven sgRNA complementary to a sequence in the human insulin promoter (PhU6-sgRNAhINS). OGM99 (5’-CACCGCGGCAGATGGCTGGGGGCTG-3’) and OGM100 (5’- pGM70 This work AAACCAGCCCCCAGCCATCTGCCGC-3’) were annealed and cloned into BbsI-digested psgRNA(MS2) (GenBank accession code MN811116).

Page 142 of 200

Plasmid Description and cloning strategy Reference

Constitutive Streptococcus pyogenes catalytically dead CRISPR associated protein 9 (dCas9) expression pGM74 vector (PhCMV-dCas9-NLS-pA). dCas9 was cloned from pGM47 (EcoRI/NheI) into pMM587 (EcoRI/SpeI) This work (GenBank accession code MN811114). pHA-NFAT1(4-460)- Constitutive expression of the regulatory domain of mouse NFAT1 (amino acids 4-460) (PhCMV-HA- 232 GFP NFAT1(4-460)-GFP-pA) (Addgene no. 11107).

233 phFLT1 Constitutive expression of human VEGFR1 (PhCMV-VEGFR1-pA) (Addgene no. 83435).

57 pHY57 PNFAT-driven Fc-stabilized shGLP1 expression vector (PNFAT-shGLP1-Fc-pA).

Constitutive Kir2.1 expression vector (PhCMV-Kir2.1-pA). Kir2.1 was excised from pGEMTEZ-Kir2.1 pKK05 Unpublished using EcoRI and cloned into the corresponding site (EcoRI) of pcDNA3.1(+).

Constitutive, PhCMV-driven expression of MCP-YPet fusion protein (PhCMV-MCP-YPet-pA). Ypet was pKK110 This work cloned from pFOX8 (SpeI/BamHI) into pMM565 (NheI/BamHI) (GenBank accession code MN811103).

Constitutive, PSV40-driven expression of a MCP-YPet fusion protein (PSV40-MCP-YPet-pA). MCP-Ypet was pKK112 cleaved from pKK110 using SpeI/BamHI and cloned into corresponding sites of pTS1016 (GenBank This work accession code MN811104). pKK122 Constitutive, PSV40-driven GEARp65 expression (PSV40-GEARp65-pA). Transcription factor p65 was PCR- This work amplified from HEK293T cDNA using OKK254 (5’-attaactagtggtggtggaGACGAACTGTTCCCCCTC-3’)

Page 143 of 200

Plasmid Description and cloning strategy Reference

and OKK255 (5’- attaggatccGCTGATCTGACTCAGCAGGG-3’), digested with SpeI/BamHI and cloned into NheI/BamHI-digested pKK112 (GenBank accession code MN811102).

PhU6-driven sgRNA complementary to a sequence in the human interleukin 2 promoter. OKK248 (5’- caccGAGGTAATGTTTTTTCAGAC-3‘) and OKK249 (5’-aaacGTCTGAAAAAACATTACCT-3‘) were pKK126 This work annealed together and cloned into the BbsI sites of the plasmid psgRNA(MS2) (GenBank accession code MN811109).

Constitutive, PSV40-driven IκB expression vector (PSV40-IκB-pA). Human IκB was PCR-amplified from HEK293T cDNA using OKK266 (5’-attaactagtTTCCAGGCGGCCGAG-3’) and OKK267 (5’- pKK130 This work attaggatccgctagcTAACGTCAGACGCTGGCCTC-3’), digested with SpeI/BamHI, and cloned into the corresponding sites of pKK112 (GenBank accession code MN811112).

Constitutive, PhCMV-driven GEARSMAD2 expression vector (PhCMV-GEARSMAD2-pA). Human SMAD2 was PCR-amplified using OKK256 (5’- attaactagtggtggtTCGTCCATCTTGCCATTC-3’) and OKK315 (5’- pKK136 This work attatctagactaTGACATGCTTGAGCAACG-3’), digested with SpeI/XbaI, and cloned into NheI/XbaI- digested pKK110 (GenBank accession code MN811108).

PhU6-driven sgRNA complementary to a sequence in the human IL-12B promoter. OKK312 (5’- pKK158 caccgAGTTTAAGTTTCCATCAGAA-3’) and OKK313 (5’-aaacTTCTGATGGAAACTTAAACTc-3’) This work were annealed and cloned into BbsI-digested psgRNA(MS2) (GenBank accession code MN811110).

Page 144 of 200

Plasmid Description and cloning strategy Reference

PhU6-driven sgRNA complementary to a sequence in the human TGFβ promoter. OKK329 (5’- pKK160 caccgCCGCCCACGCGAGATGAGGA-3’) and OKK330 (5’-aaacTCCTCATCTCGCGTGGGCGGC-3’) This work were annealed and cloned into BbsI-digested psgRNA(MS2) (GenBank accession code MN811111).

ITR-containing vector for SB100X-specific transposon mediated stable genomic integrations, containing a constitutive dTomato and BlastR expression unit, and a constitutive expression unit for Streptococcus

pyogenes catalytically dead CRISPR associated protein 9 (dCas9) (ITR-PhEF1α-dCas9-NLS-pA:PRPBSA- pKK171 dTomato-P2A-BlastR-pA-ITR). dCas9-NLS was PCR-amplified from pGM74 using OKK183 (5’- attaggcctgacaggccTTAGGGGTCCTCCACCTTGCGCTTCTTCTTGGGATCCCCTCCGAGCTGTG-3‘) and OKK184 (5’-attaggcctctgaggccACCATGACTAGTGACAAGAAGTATTCTATCGG-3‘), cleaved with SfiI and cloned into the corresponding site of pSBbi-RB (GenBank accession code MN811107).

ITR-containing vector for SB100X-specific transposon mediated stable genomic integrations, containing a

constitutive dTomato and BlastR expression unit, and a constitutive expression unit for GEARNFAT (ITR-

PhEF1α-GEARNFAT-pA:PRPBSA-dTomato-P2A-BlastR-pA-ITR). GEARNFAT was PCR-amplified from pKK172 This work pKK50 using OKK350 (5’-attaggcctctgaggccACCATGGCTTCAAACTTTACTCAGTTCG-3‘) and OKK351 (5’-attaggcctgacaggccTTAAACGGGCCCTCTAGACTC-3‘), cleaved with SfiI and cloned into corresponding site of pSBbi-RB (GenBank accession code MN811106). pKK44 Constitutive expression of a fusion protein that consists of MCP, the transactivation domain of transcription This work factor p65 (p65TA), and the transactivation domain of heat shock factor 1 (HSF1TA), without nuclear

Page 145 of 200

Plasmid Description and cloning strategy Reference

localization signal (PhCMV-MCP-p65TA-HSF1TA-pA). MCP was PCR-amplified with OKK84 (5’- attaaagcttCGTACGGCCACCATGG-3’) and OKK85 (5’-CGCTACCTCCTCCTCCGCTTC-3’) from

pMS2-P65-HSF1_Hygro. p65TA-HSF1TA was PCR-amplified with OKK86 (5’- AAGCGGAGGAGGAGGTAGCGGACCTTCAGGGCAGATCAGC-3’) and OKK87 (5’- attagaattcTGTACAGGAGACAGTGGGGTC-3’). The resulting fragments were assembled with a PCR reaction, amplified using OKK84 and OKK87, digested with HindIII/EcoRI, and cloned into corresponding sites of pcDNA3.1(+) (GenBank accession code MN811105).

Constitutive GEARNFAT expression vector (PhCMV-GEARNFAT-pA). MCP-p65TA-HSF1TA was PCR- amplified from pKK44 using OKK84 (5’-attaaagcttCGTACGGCCACCATGG-3’) and OKK96 (5’-

CGCTGCCTCCTGAACCGCCGCTTCCGCCTGTACAGGAGACAGTGGGGTC-3’). NFATreg-GFP was PCR-amplified from pHA-NFAT1(4-460)-GFP using OKK97 (5’- pKK50 GCGGTTCAGGAGGCAGCGGTGGATCAGGCTCCACCATGATCTTTTACCC-3’) and OKK93 (5’- This work

attagaattcCTTGTACAGCTCGTCCATGC-3‘). The resulting fragments were assembled by PCR to create

MCP-p65TA-HSF1TA-NFATreg-GFP, were amplified using OKK84 and OKK93, and were cloned into pcDNA3.1(+) using HindIII/EcoRI (GenBank accession code MN811100).

Constitutive GEARElk expression vector. (PhCMV-GEARElk1-pA) MCP was PCR-amplified with OKK155 pKK82 (5’-attaacgcgtaagcttCGTACGGCCACCATG-3’) and OKK85 (5’-CGCTACCTCCTCCTCCGCTTC-3‘) This work from pKK44. Elk1 was PCR-amplified from pCGN-ELK-1 using OKK128 (5’-

Page 146 of 200

Plasmid Description and cloning strategy Reference

gaggaagcggaggaggaggtagcggaccGATCTCCCAGCCGCAG-3’) and OKK129 (5’- ctagaaggcacagtcgaggctgatctagaTCATGGCTTCTGGGG-3’). The resulting fragments were assembled by PCR to create MCP-Elk1, were amplified using OKK155 and OKK129, and were cloned using HindIII/XbaI into pcDNA3.1(+) (GenBank accession code MN811101).

PNFAT-driven secreted stabilized NanoLuc expression vector (PNFAT–NanoLuc-Fc-pA). NanoLuc was PCR- amplified from pProinsulin-NanoLuc using OKK193 (5’- attagaattcgccaccATGGGCGTCAAGGTGCTGTTCGCCCTCATTTGTATAGCTGTCGCTGAGGCGgtct tcacactcgaagatttcg-3’) and OKK194 (5’-CCACTTCCACCGCCTCCCGCCAGAATGCGTTCG-3’). Fc pKK89 fragment of a murine IgG was PCR-amplified from pHY57 using OKK191 (5’- This work GGAGGCGGTGGAAGTGGTGGTTGTAAGCCTTGCATATGTAC-‘3’) and OKK180 (5’- attagtcgacCCACATTTGTAGAGGTTTTACTTGC-3’). The resulting fragments were assembled by PCR using OKK193 and OKK180, cleaved with EcoRI/SalI and cloned into corresponding sites of pMX57 (GenBank accession code MN811113).

ITR containing vector for SB100X-specific transposon mediated stable genomic integrations, containing a constitutive dTomato, PuroR expression unit and a constitutive expression unit for sgRNA complementary pLeo1207 This work to a sequence in the human insulin promoter (ITR-PhU6-sgRNAhINS:PRPBSA-BFP-P2A-PuroR-pA-ITR). PhU6-

sgRNAhINS was PCR-amplified from pGM70 using OKK356 (5’- attaaccggtCGAGGGCCTATTTCCCATGATTCCTTC-3‘) and OKK357 (5’-

Page 147 of 200

Plasmid Description and cloning strategy Reference

attatctagaCAGATGCGTAAGGAGAAAATACCGCATC-3‘), cleaved with AgeI and XbaI and cloned into corresponding sites of pSBbi-BP (GenBank accession code MN811118).

ITR containing vector for SB100X-specific transposon mediated stable genomic integrations, containing a constitutive dTomato, PuroR expression unit and a constitutive expression unit for sgRNA complementary

to a sequence in the human interleukin 12B promoter (ITR-PhU6-sgRNAhIL-12B:PRPBSA-BFP-P2A-PuroR-pA- pLeo1209 ITR). PhU6-sgRNAhIL-12B was PCR-amplified from pKK158 using OKK356 (5’- This work attaaccggtCGAGGGCCTATTTCCCATGATTCCTTC-3‘) and OKK357 (5’- attatctagaCAGATGCGTAAGGAGAAAATACCGCATC-3‘), cleaved with AgeI and XbaI and cloned into corresponding sites of pSBbi-BP (GenBank accession code MN811119). pLeo628 Constitutive expression of MAPK-GEMSRR120 (PSV40-MAPK-GEMSRR120 -pA) This work

234 pMF111 Mammalian reporter plasmid for TetR-Elk1 induced SEAP expression (OTetR-PhCMVmin-SEAP-pA)

141 pMM1 Cloning vector for mammalian gene expression (PhCMV-MCS-pA).

235 pMM565 Constitutive expression of MCP (PhCMV-MCP-pA).

236 pMM587 Constitutive expression of an NLS peptide (PhCMV-NLS-pA).

Page 148 of 200

Plasmid Description and cloning strategy Reference

Lentiviral vector containing a fusion protein consisting of MCP, a nuclear localization signal, the pMS2-P65- transactivation domain of transcription factor p65 (p65TA), and the transactivation domain of heat shock Unpublished HSF1_Hygro factor 1 (HSF1TA) (PEF1α-MCP-NLS-p65TA- HSF1TA-pA) (Addgene no. 61426).

188 pMX57 PNFAT-driven SEAP expression vector (PNFAT-SEAP-pA)

pProinsulin-NanoLuc Lentiviral vector for constitutive expression of Proinsulin-NanoLuc (PhCMV-Proinsulin-NanoLuc-pA). 46 (Addgene no. 62057). pSBbi-BP ITR-containing vector for SB100X-specific transposon-mediated stable genomic integrations, containing a 168 constitutive BFP and PuroR expression unit, and a constitutive expression unit (ITR-PhEF1α-MCS-

pA:PRPBSA-BFP-P2A-PuroR-pA-ITR). (Addgene no. 60512) pSBbi-RB ITR-containing vector for SB100X-specific transposon-mediated stable genomic integrations, containing a 189 constitutive dTomato and BlastR expression unit, and a constitutive expression unit (ITR-PhEF1α-MCS-

pA:PRPBSA-dTomato-P2A-BlastR-pA-ITR). (Addgene no. 60512) psgRNA(MS2) sgRNA cloning vector with MCP-binding loops at tetraloop and stemloop 2 (Addgene no. 61424). 189

188 pSP20 PhINS-driven SEAP expression vector (PhINS-SEAP-pA) (GenBank accession code MN811117).

Page 149 of 200

Plasmid Description and cloning strategy Reference pTBRII Constitutive TBRII expression vector (PhCMV-TBRII-pA) (Addgene no. 16622). This work

237 pTS1016 Constitutive, PSV40-driven SEAP expression (PSV40-SEAP-pA).

Constitutive, PhCMV-driven vector expressing MESA target chain with V2-MESA ectodomain, 35 pV1-MESA-35F-M- extracellular linkers, a flag tag, the methionine-containing cleavage sequence, and tTA (PhCMV-MESA V2 Unpublished tTA tTA-pA). (Addgene no. 84502).

Constitutive, PhCMV-driven vector expressing MESA target chain with V1-MESA ectodomain, 45 pV1-MESA-45F-M- 202 extracellular linkers, a flag tag, the methionine-containing cleavage sequence, and dCas9-VP64 (PhCMV- dCas9 MESA V1 dCas9-VP64-pA). (Addgene no. 84504).

Constitutive, PhCMV-driven vector expressing MESA target chain with V1-MESA ectodomain, 45 pV1-MESA-45F-M- 202 extracellular linkers, a flag tag, the methionine-containing cleavage sequence, and tTA (PhCMV-MESA V1 tTA tTA-pA). (Addgene no. 84500).

Constitutive, PhCMV-driven vector expressing MESA protease chain with V1-MESA ectodomain, 45 pV1-MESA-45F-Tev 202 extracellular linkers, a flag tag, and Tev protease (PhCMV-MESA V1 TEV-pA). (Addgene no. 84501).

Constitutive, PhCMV-driven vector expressing MESA target chain with V2-MESA ectodomain, 35 pV2-MESA-35F-M- 202 extracellular linkers, a flag tag, the methionine-containing cleavage sequence, and dCas9-VP64 (PhCMV- dCas9 MESA V2 dCas9-VP64-pA). (Addgene no. 84506).

Page 150 of 200

Plasmid Description and cloning strategy Reference

Constitutive, PhCMV-driven vector expressing MESA protease chain with V2-MESA ectodomain, 35 pV2-MESA-35F-Tev 202 extracellular linkers, a flag tag, and Tev protease (PhCMV-MESA V2 TEV-pA). (Addgene no. 84503).

Oligonucleotides: Restriction endonuclease-specific sites are shown in italics and annealing sequences are shown in capital letters.

Abbreviations: α1D, α 1D subunit of the murine L-type voltage-gated calcium channel Cav1.3; α2/δ1, α2 and δ1 subunits of the murine L-type voltage- gated calcium channel; β3, β3 subunit of the murine L-type voltage-gated calcium channel; BFP, blue fluorescent protein; BlastR, blasticidin resistance gene; Cav1.3, member 3 of the Cav1 family of L-type voltage-gated calcium channels; CaRROT, calcium-responsive transcriptional reprogramming tool; CRISPR, clustered regularly interspaced short palindromic repeats; dCas9, catalytically inactive Streptococcus pyogenes CRISPR associated protein 9; dTomato, dimeric red fluorescent protein variant; Fc – Fc fragment of murine IgG antibody; HEK293T, human endothelial kidney 293 cell line with stably incorporated Simian virus large T antigen; HSF1, human heat shock factor 1; HSF1TA, human heat shock factor 1 transactivation domain; GEAR, generalized engineered activation regulator; GEARElk, GEAR containing the transactivation domain of human Elk1; GEARNFAT, GEAR containing the regulatory domain of murine NFAT1; GEARp65, GEAR containing the human transcription factor p65; GEARSMAD2, GEAR containing the human

SMAD2; MAPK-GEMSRR120, generalized extracellular molecule sensor activating MAPK pathway specific to RR120; GFP, green fluorescent protein; HygroR, hygromycin resistance gene; ITR, inverted terminal repeats of SB100X; IκB, nuclear factor of kappa light polypeptide gene enhancer in B- cells inhibitor; MAPK, mitogen-activated protein kinase; MCP, MS2 bacteriophage coat protein; MCS, multiple cloning site; MESA, modular extracellular sensor; NanoLuc, Oplophorus gracilirostris luciferase; NLS, nuclear localization signal; NFAT, nuclear factor of activated T-cells;

NFATreg, nuclear factor of activated T-cells regulatory domain; OTetR, TetR-binding operator sequence; p65, human transcription factor p65; p65TA, human transcription factor p65 transactivation domain; pA, polyadenylation signal; PCR, polymerase chain reaction; PhCMV, human cytomegalovirus immediate early promoter; PhEF1α, human elongation factor 1 alpha promoter; PhINS, human insulin promoter; PNFAT, synthetic mammalian promoter

Page 151 of 200

containing five tandem repeats of a human IL-4 NFAT-binding site; PRPBSA, constitutive synthetic mammalian promoter; PSV40, simian virus 40 promoter; PuroR, puromycin resistance gene; SB100X, optimized Sleeping Beauty transposase; SEAP, human placental secreted alkaline phosphatase; sgRNA, synthetic guide RNA; shGLP1, short human glucagon-like peptide 1; SMAD2, mothers against decapentaplegic homolog 2; TBRII, transforming growth factor β type II receptor; TEV, tobacco etch virus nuclear-inclusion-a endopeptidase; tTA, tetracycline-controlled transactivator; VEGFR1, vascular endothelial growth factor receptor, type 1; YPet, yellow fluorescent protein variant.

Page 152 of 200

Supplementary Table 2. Synthetic guide RNA target sequences Plasmid Target Complementary sequence Reference

pGM70, pLeo1207 Human insulin promoter 5’-CGGCAGATGGCTGGGGGCTG-3’ This work

pKK126 Human interleukin 2 (IL-2) promoter 5’-AGGTAATGTTTTTTCAGAC-3’ 226

Human interleukin 12 B (IL-12B) pKK158, pLeo1209 5’- AGTTTAAGTTTCCATCAGAA-3’ This work promoter

pKK160 Human TGFβ promoter 5’- CCGCCCACGCGAGATGAGGA-3’ This work

Page 153 of 200

Supplementary Table 3. Detailed transfection protocols used for Figures 2, 3, 5 and Supplementary Figures 3 and 9.

The amount of DNA was calculated per 1 well of a 24-well plate. Application GEAR sgRNA Other plasmids Total DNA

pGM74 – 25 ng

pKK05 – 15 ng

Membrane depolarization- pKK50 pGM70 pSP20 – 150 ng 300 ng induced transgene activation 10 ng 25 ng pCaV1.3 – 30 ng

pCaVα2δ1 – 30 ng

pCaVb3 – 10 ng

pGM74 – 50 ng

pKK05 – 15 ng

Membrane depolarization- pKK50 pGM70 pCOLADuet-1 – 130 ng induced endogenous insulin 300 ng activation 10 ng 25 ng pCaV1.3 – 30 ng

pCaVα2δ1 – 30 ng

pCaVb3 – 10 ng

Page 154 of 200

Application GEAR sgRNA Other plasmids Total DNA

pGM74 – 50 ng

VEGF165-induced transgene pKK50 pGM70 pSP20 – 150 ng 300 ng activation 2.5 ng 25 ng phFLT1 – 25 ng

pCOLADuet-1 – 47.5 ng

pGM74 – 50 ng VEGF165-induced pKK50 pGM70 phFLT1 – 25 ng 300 ng endogenous gene activation 6.25 ng 25 ng pCOLADuet-1 – 193.75 ng

pGM74 – 50 ng

TGFβ-induced transgene pKK136 pGM70 pSP20 – 150 ng 300 ng activation 25 ng 25 ng pTBRII – 25 ng

pCOLADuet-1 – 25 g

TGFβ-induced endogenous pKK136 pGM70 or pKK158 pGM74 – 50 ng 300 ng gene activation 25 ng 25 ng pTBRII – 25 ng

Page 155 of 200

Application GEAR sgRNA Other plasmids Total DNA

pCOLADuet-1 – 175 ng

pGM74 – 25 ng TNFα-induced transgene pKK122 pGM70 pSP20 – 150 ng 500 ng activation 150 ng 25 ng pKK130 – 150 ng

TNFα-induced endogenous pKK122 pGM70 or pKK158 pGM74 – 50 ng 475 ng gene activation 200 ng 25 ng pKK130 – 200 ng

pGM74 – 50 ng bFGF-induced transgene pKK82 pGM70 pSP20 – 150 ng 300 ng activation 25 ng 25 ng pCOLADuet-1 – 50 ng

pGM74 – 25 ng RR120-induced transgene pKK82 pGM70 pSP20 – 150 ng 500 ng activation 25 ng 25 ng pLeo628 – 275 ng

Supplementary Table 4. qPCR primer pairs

Page 156 of 200

Target Sequence Reference

5‘-ATCAGAAGAGGCCATCAAGCA-3‘ Human insulin 238 5‘-TAGAGAGCTTCCACCAGGTGTGA-3‘

5‘-GTCTCCTCTGACTTCAACAGCG-3‘ Origene Technologies Human GAPDH 5‘-ACCACCCTGTTGCTGTAGCCAA-3‘ (Cat. No. HP205798)

5‘-GACATTCTGCGTTCAGGTCCAG-3‘ Origene Technologies Human interleukin 12B 5‘-CATTTTTGCGGCAGATGACCGTG-3‘ (Cat. no. HP205923)

5‘-ACAAACTGGGGCCTGAGATACC-3‘ SEAP This work 5‘-CTGCACTCAAGCCAATGGTCTG-3‘

5‘-TACCTGAACCCGTGTTGCTCTC-’3 Origene Technologies Human TGFβ 5‘-GTTGCTGAGGTATCGCCAGGAA-3’ (Cat. no. HP200609)

5‘-AGAACTCAAACCTCTGGAGGAAG-’3 Origene Technologies Human interleukin 2 5‘-GCTGTCTCATCAGCATATTCACAC-’3 (Cat. no. HP200553)

Human nuclear receptor related-1 protein 5‘-AAACTGCCCAGTGGACAAGCGT-’3 Origene Technologies (NR4A2) 5‘-GCTCTTCGGTTTCGAGGGCAAA-’3 (Cat. no. HP209171)

Human early growth response 3 (EGR3) 5‘-GACTCGGTAGTCCATTACAATCAG-’3 Origene Technologies

Page 157 of 200

Target Sequence Reference

5‘-AGTAGGTCACGGTCTTGTTGCC-’3 (Cat. no. HP207741)

Human vascular cell adhesion molecule 1 5‘-GATTCTGTGCCCACAGTAAGGC-’3 Origene Technologies (VCAM1) 5‘-TGGTCACAGAGCCACCTTCTTG-’3 (Cat. no. HP230503)

Page 158 of 200

Supplementary Table 5. Detailed statistics

Confidence level = 95%

F test to compare Welch’s correction Two-tailed Mean 1 Mean 2 Fold Figure Description variances applied t-test (control) (induced) induction (p-value) (yes/no) (p-value)

2b CaV1.3, GEARNFAT; 3.71 33.5 9.0 0.0092 yes 0.0063

2d VEGFR, GEARNFAT 6.54 13.27 2.0 0.090 no 0.0035

TBRII, 2f 7.71 50.76 6.6 0.33 no 0.00014 GEARSMAD2

TNFR, 2h 4.55 89.28 19.6 0.0067 yes 0.0085 GEARp65

FGFR, 2j 2.10 9.54 4.5 0.69 no 0.00000037 GEARElk1

GEMS-MAPK (RR120), 2l 1.40 14.41 10.3 0.067 no 0.0000057 GEARElk1

Ca 1.3, GEAR – Insulin 3b V NFAT 3.02 97.42 32.3 0.0010 yes 0.011 mRNA

Page 159 of 200

F test to compare Welch’s correction Two-tailed Mean 1 Mean 2 Fold Figure Description variances applied t-test (control) (induced) induction (p-value) (yes/no) (p-value)

Ca 1.3, GEAR - Insulin 3c V NFAT 0 0.26 N/A N/A yes 0.0004 protein

Insulin 6.77 234.5 34.6 0.0097 yes 0.0027 3d IL-12 11.3 81.26 7.2 0.017 yes 0.021

gRNAIL-12 12.34 8546 692.5 0.00000065 yes 0.032 4b gRNAINS 2.87 3.17 1.1 0.16 no 0.7314

2 h 18.37 25.7 1.4 0.21 no 0.17

4c 4 h 23.92 322.3 13.5 0.037 yes 0.0008

8 h 20.93 2112 100.9 0.74 no 0.00000000004

TBRII, 5b 8.83 14.51 1.6 0.48 no 0.0080 GEARSMAD2

TNFR, 5d 20.04 193.8 9.7 0.044 yes 0.020 GEARp65

S2 8 h 0.84 3.002 3.6 0.020 yes 0.0397

Page 160 of 200

F test to compare Welch’s correction Two-tailed Mean 1 Mean 2 Fold Figure Description variances applied t-test (control) (induced) induction (p-value) (yes/no) (p-value)

24 h 6.80 30.16 4.4 0.08 no 0.000006

36 h 9.731 49.26 5.1 0.0045 yes 0.0019

48 h 48.2 117.8 2.4 0.8037 no 0.0003

S3a VEGFR, GEARNFAT 14.2 29.23 2.1 0.95 no 0.014

TBRII, S3b 0 87.31 N/A N/A yes 0.018 GEARSMAD2

TNFR, S3c 4.30 39.3 9.2 0.29 no 0.000082 GEARp65

S4a VEGFR/GEARNFAT 0.74 2.42 3.3 0.39 no 0.0164

S4b TNFR/ GEARp65 0.098 2.19 22.4 0.014 yes 0.0305

S4c TBRII/ GEARSMAD2 0.11 0.65 5.7 0.12 no 0.0008

S4d FGFR/GEARElk1 0.087 0.29 3.3 0.73 no 0.0314

MESA V1 TetR 24 h 20 ng 19.65 20 1.0 0.82 no 0.825 S5a MESA V1 TetR 36 h 20 ng 44.72 47.85 1.1 0.498 no 0.073

Page 161 of 200

F test to compare Welch’s correction Two-tailed Mean 1 Mean 2 Fold Figure Description variances applied t-test (control) (induced) induction (p-value) (yes/no) (p-value)

MESA V1 TetR 24 h 200 ng 19.65 17.51 0.9 0.59 no 0.237

MESA V1 TetR 36 h 200 ng 44.72 48.5 1.1 0.0045 no 0.063

MESA V1 TetR 24 h 20 ng 0.82 0.97 1.2 0.72 no 0.107

MESA V1 TetR 36 h 20 ng 1.65 1.98 1.2 0.33 no 0.005 S5b MESA V1 TetR 24 h 200 ng 0.82 0.80 1.0 0.19 no 0.839

MESA V1 TetR 36 h 200 ng 1.65 1.96 1.2 0.34 no 0.013

MESA V1 dCas9 24 h 20 ng 0.55 0.50 0.9 0.79 no 0.579

MESA V1 dCas9 36 h 20 ng 1.61 1.62 1.0 0.17 no 0.902

MESA V1 dCas9 24 h 200 0.115 S5c 0.55 0.39 0.7 0.70 no ng

MESA V1 dCas9 36 h 200 0.021 1.61 1.26 0.8 0.18 no ng

MESA V1 dCas9 24 h 20 ng 1.61 1.62 1.0 0.17 no 0.948 S5d MESA V1 dCas9 36 h 20 ng 2.56 2.66 1.0 0.43 no 0.569

Page 162 of 200

F test to compare Welch’s correction Two-tailed Mean 1 Mean 2 Fold Figure Description variances applied t-test (control) (induced) induction (p-value) (yes/no) (p-value)

MESA V1 dCas9 24 h 200 0.021 1.61 1.26 0.8 0.18 no ng

MESA V1 dCas9 36 h 200 0.299 2.56 2.36 0.9 0.58 no ng

MESA V2 TetR 24 h 20 ng 16.6 18.14 1.1 0.45 no 0.133

MESA V2 TetR 36 h 20 ng 67.64 71.71 1.1 0.11 no 0.232 S5e MESA V2 TetR 24 h 200 ng 16.6 16.49 1.0 0.89 no 0.924

MESA V2 TetR 36 h 200 ng 67.64 69.8 1.0 0.70 no 0.569

MESA V2 TetR 24 h 20 ng 0.98 1.16 1.2 0.0032 yes 0.128

MESA V2 TetR 36 h 20 ng 2.42 2.52 1.0 0.34 no 0.570 S5f MESA V2 TetR 24 h 200 ng 0.98 1.06 1.1 0.0008 no 0.681

MESA V2 TetR 36 h 200 ng 2.42 2.52 1.0 0.79 no 0.405

MESA V2 dCas9 24 h 20 ng 0.11 0.097 0.9 0.39 no 0.819 S5g MESA V2 dCas9 36 h 20 ng 0.15 0.091 0.6 0.75 no 0.175

Page 163 of 200

F test to compare Welch’s correction Two-tailed Mean 1 Mean 2 Fold Figure Description variances applied t-test (control) (induced) induction (p-value) (yes/no) (p-value)

MESA V2 dCas9 24 h 200 0.981 0.11 0.11 1.0 0.43 no ng

MESA V2 dCas9 36 h 200 0.033 0.15 0.059 0.4 0.028 yes ng

MESA V2 dCas9 24 h 20 ng 0.24 0.30 1.3 0.34 no 0.250

MESA V2 dCas9 36 h 20 ng 0.49 0.63 1.3 0.54 no 0.353

MESA V2 dCas9 24 h 200 0.831 S5h 0.24 0.25 1.1 0.19 no ng

MESA V2 dCas9 36 h 200 0.839 0.49 0.52 1.1 0.05 no ng

IL-12 0.00 1.47 N/A N/A yes 0.50

IL-2 1301.47 3948.24 3.0 0.46 no 0.0033

S6a NR4A2 0.83 642.07 770.7 0.32 no 0.00000006

EGR3 85.77 1072.99 12.5 0.063 no 0.00092

FOSb 0.87 2.75 3.2 0.24 no 0.48

Page 164 of 200

F test to compare Welch’s correction Two-tailed Mean 1 Mean 2 Fold Figure Description variances applied t-test (control) (induced) induction (p-value) (yes/no) (p-value)

ATF3 401.20 322.81 0.8 0.51 no 0.30

VCAM1 0.00 342.29 N/A N/A yes 0.037

IL-12 3.76 226.73 60.3 0.0042 yes 0.0073

IL-2 593.28 1012.55 1.7 0.18 no 0.11

NR4A2 2.52 331.47 131.4 9.15E-05 yes 0.027

S6b EGR3 52.70 358.27 6.8 0.020 yes 0.0007

FOSb 2.07 13.18 6.4 0.041 yes 0.091

ATF3 203.23 456.59 2.2 0.066 no 0.020

VCAM1 0.00 16.63 N/A N/A yes 0.082

GEARNFAT 1.389 23.63 11.0 0.018 yes 0.0018 S8a CaRROT 1.61 5.55 3.4 0.82 no 0.00002

24 h 1.36 2.25 1.7 0.050 no 0.0021 S8b 48 h 2.19 5.31 2.4 0.058 no 0.0010

S9a GEARNFAT 12.47 82.2 6.6 0.98 no 0.0000015

Page 165 of 200

F test to compare Welch’s correction Two-tailed Mean 1 Mean 2 Fold Figure Description variances applied t-test (control) (induced) induction (p-value) (yes/no) (p-value)

No GEAR 0.52 1.80 3.5 0.077 no 0.00014 control

No gRNA 0.73 1.98 2.7 0.67 no 0.0032 control

no GEARNFAT 5.235 10.03 1.9 0.18 0.0010

No GEAR S9b 1.00 1.29 1.3 0.92 no 0.14 control

No gRNA 1.29 1.34 1.0 0.13 no 0.93 control

GEARSMAD2 7.87 77.25 9.8 0.063 no 0.00023

No GEAR S9c 0.71 1.02 1.4 0.11 no 0.50 control

No gRNA 0.94 1.00 1.1 0.15 no 0.58

Page 166 of 200

F test to compare Welch’s correction Two-tailed Mean 1 Mean 2 Fold Figure Description variances applied t-test (control) (induced) induction (p-value) (yes/no) (p-value)

control

GEARp65 16.89 215.8 12.8 0.021 yes 0.0087

No GEAR 0.80 3.74 4.6 0.75 no 0.000036 S9d control

No gRNA 3.16 6.42 2.0 0.22 no 0.0049 control

GEARElk1 2.51 11.47 4.6 0.37 no 0.0000082

No GEAR 0.88 0.62 0.7 0.77 no 0.14 S9e control

No gRNA 5.82 7.19 1.2 0.064 no 0.019 control

GEARNFAT 0.61 117.58 192.74 0.000050 yes 0.00059

S10a No GEAR N/A N/A N/A N/A N/A N/A control

Page 167 of 200

F test to compare Welch’s correction Two-tailed Mean 1 Mean 2 Fold Figure Description variances applied t-test (control) (induced) induction (p-value) (yes/no) (p-value)

No gRNA N/A N/A N/A N/A N/A N/A control

GEARNFAT 4.88 26.68 5.87 0.13 no 0.0034

No GEAR 3.36 4.76 1.41 0.08 no 0.11 S10b control

No gRNA 7.70 14.75 1.92 0.61 no 0.038 control

S13a SEAP mRNA 0.88 7.43 8.5 0.0316 yes 0.045

S13b SEAP protein 17.16 114.5 6.7 0.0070 yes 0.018

N/A – not applicable

Page 168 of 200

Discussion

In electronic engineering one can distinguish hardware (physical components), and software (a set of instructions). Interdisciplinary bioengineering involves also a third element, bioware, which refers to biological components of the system. This work focused on developing bioware itself, as well as on communicating it with electronic hardware. Synthetic biology- inspired therapies could benefit from new ways to regulate the behavior of therapeutic cells provided by electrogenetics and GEARs.

Electrogenetic cellular insulin release for real-time glycemic control in type-1 diabetic mice

Rapid progress of electronics resulted in the development of numerous medical technologies. Electronic wearables can not only read basic body parameters like movement, heart rate, glucose concentration in tears or saliva239, but also integrate them into complex biometric data: sleep duration and quality, respiratory rate, heart rate variability, ECG, posture, and body temperature changes240-243. In the future, we can expect personal vapor sensors detecting disease biomarkers in breath244, and wristbands estimating the user’s emotional state245, or even a direct connection to a nervous system246,247. The concept of using electronic implants to break the boundaries imposed on humans by their biology inspired scientists and science fiction writers for decades. From the electronic engineering point of view, building an implant equipped with a microcontroller, capable of communicating with computers, smartphones, or sensors is already possible. Wireless charging, which currently gains popularity, could eliminate the necessity of removing the implant. Hence, one could imagine a compact, autonomous implant-and-forget device, which would automatically regulate the host’s physiology. However, constructing an implant capable of electronically controlled secretion of a therapeutic agent requires designing an interface between engineered therapeutic cells and an electronic device. Consequently, it requires finding a stimulus, which could be easily generated by an electronic device and sensed by a living cell.

Heat was the first physical stimulus that was used to control engineered cells via heat shock factors. However, relying on a mechanism that evolved to protect cells from damage and becomes activated when that damage occurs is problematic, because of the narrow window between activation of a transgene expression and the cell’s death. Heat-sensitive ion channels that are involved in the temperature reception by sensory neurons enable cell activation in lower

Page 169 of 200

temperatures. Yet, thermal energy transfer for in vivo applications remains a challenge. Magneto- and radiogenetics solved that problem by using an electromagnetic field to heat up nanoparticles coupled to thermosensitive ion channels. Unfortunately, metal nanoparticles cannot be synthetized by the engineered cells, which limits their applicability in vivo. Although it was reported that heat transfer via genetically encoded ferritin nanoparticles could activate transgene expression, some researchers claimed that such systems would violate laws of physics and cannot function as reported 100,101. Additionally, heat-sensitive systems could be affected by thermoregulation and by fluctuations in body temperature. Optogenetics still celebrates its golden age in synthetic biology. Electric current converted to light was already used to induce transgene expression in a biomedical implant and to control diabetes16,17. Although that technology has a great potential for further improvements, optogenetic systems face multiple challenges, which include: (i) wavelength-dependent cytotoxicity of light150, (ii) the use of bacterial components that could stimulate immune response of the host16,17,144-146,248, (iii) the use of pathways related to viral infection response which could interact with the host’s immune reaction, or (iv) the need for inorganic cofactors that cannot be easily synthetized by the therapeutic cell and that have side effects 151-153, poor bioavailability or short half-lives in vivo 154. Furthermore, the energy consumption of systems that were used in the previous designs of optogenetic implants was high, as cells required long light pulses to get activated16,17,249,250.

This work presents the first truly direct interface between electronic devices and engineered therapeutic cells. It fully relies on mammalian components that can be produced by the cells, does not require inorganic cofactors, and does not involve pathways responsible for detecting viral infections. Eliminating intermediate energy forms reduces losses on energy transition, which makes electrogenetics suitable for constructing autonomous, battery-powered implants that could theoretically operate for many months without charging. Furthermore, electrogenetic technology is compatible with natural secretory mechanisms, which allows for bypassing transcriptional delay typical for previously published synthetic biology-inspired therapies.

Slow kinetics is inherently associated with transcriptionally controlled systems, which need hours to respond. The regulation of glucose levels in the blood, a basic homeostatic process required for maintaining a proper energy supply for the body, is controlled by rapid vesicular secretion of two hormones: insulin and glucagon. Because the vesicular secretion of both hormones depends on membrane depolarization, it can be triggered by electrical pulse

Page 170 of 200

stimulation. An engineered therapeutic cell line Electroβ is capable of insulin secretion upon electrical pulse stimulation within minutes. Combining it with a bioelectronic implant allows for real-time glycemic control in diabetic mice. Similar fast release upon electrical stimulation was shown in vitro with glucagon-secreting cells; thereby proving that electrogenetics is not limited to a single application.

Being a newly established field, electrogenetics has a large potential for improvement. Decreasing stimulation voltage and increasing energetic efficiency of the interface would be one of the major engineering milestones. In this study, two methods of electrical stimulation were presented. Switching from free-hanging electrodes to trans-layer stimulation decreased the voltage by an order of magnitude, to 7.5-10 V. Further reducing that value to 3.3-5 V, which is the standard voltage of lithium-ion batteries and electronic devices, would increase the compatibility of electrogenetics with standard electronics. Furthermore, such change would allow simplifying the electronic circuit of a bioelectronic implant and save energy - a voltage- amplifying step would become redundant if a stimulation impulse could come directly from a microcontroller. Optimally, the stimulation voltage should be below 1.23 V, which is the threshold value for water electrolysis251. Possible improvements can be made at the level of cell engineering, or the interface design. Naturally excitable cells, like neurons or cardiomyocytes, evolved a function-specific structure, such as a prolonged shape, internal compartmentalization, or an ion channel-based signal amplifier252,253. Implementing such modifications into engineered electrosensitive cells could increase their sensitivity to electrostimulation. Despite the electrical cell stimulation was widely studied, there are still controversies regarding the exact mechanism. Two common theories assume an interaction between the electric field and the transmembrane field254, or a redistribution of charge inside the cell255. Regardless of the theoretical principle, it was shown that the positioning of cells in a field, and therefore, an interface design, affects the threshold strength of the stimulation256,257. It is worth to mention that multielectrode array matrices (MEAs), commonly used for studying neurons, typically require less than 1 V to trigger action potential258,259.

As the effect of the electric field on voltage-gated ion channels depends on its local gradient, reducing the distance between two electrodes alone can be sufficient to decrease the stimulation voltage. A change from the free-hanging electrodes setup to the transmembrane stimulation performed in this work led to a fivefold decrease of the voltage necessary to activate

ElectroHEK cells. This change was associated with a reduction of the distance between the

Page 171 of 200

electrodes from approximately 25 mm to approximately 5 mm. However, optimizing the electrode positioning requires not only maximizing the voltage-to-distance ratio, but also ensuring the current flow through the desired path. An inherent property of the electric current is its concentration in the path of the least resistance (according to the Ohm’s law). Consequently, simply decreasing the distance between two electrodes immersed in a cell- containing conductive medium resulted in lower stimulation efficiency (data not shown). In the transmembrane stimulation setup, a porous membrane covered by cells has higher electrical resistance than the cell culture medium, or the extracellular fluid. Such solution facilitates more uniform field distribution, and forces the current flow through the cells. MEAs, on the other hand, rely on the stimulation of single cells by microelectrodes. Future designs of the electrogenetic interface could enable more efficient stimulation with lower energy requirements by optimizing the voltage gradient, electrode positioning and the current flow path.

The energy source for a bioelectronic implant needs to deliver a sufficient amount of current and voltage for a certain period of time. Wireless energy transfer simplifies the electronic design, requires less space, and does not require regular recharging. Furthermore, it completely mitigates the risk of leakage or exposition related to batteries. Although such solution was chosen for the implant presented in this work, a battery-powered circuit offers other advantages that make them attractive subjects of future research. Especially, an internal power source would allow the construction of an autonomous, smart implant that could communicate with other electronic devices, for example to establish a closed-loop communication with a glucose sensor.

Therapeutic cells intended for use in humans do not only need to be effective and robust, but also safe for the patient. Future work towards clinical applications would be required to ensure that electrogenetic cells secrete only the intended product, remain free from adventitious agents and do not cause excessive immune response. Although cell encapsulation introduces a mechanical barrier that limits their spreading in the host, it would be desired to develop a non- cancerous cell line with contact inhibition of proliferation. Besides increased safety, such cells would reduce the amount of energy spent on clinically non-relevant metabolic processes, which could potentially decrease the consumption of nutrients and oxygen, production of metabolites, and therefore increase their lifetime inside of the implant. Because encapsulated cells are protected from cell-mediated immune response, they do not need to be autologous.

Page 172 of 200

Consequently, engineering a generic, off-the-shelf therapeutic cell line would greatly reduce the cost of treatment and increase its availability.

Long-term studies revealed limited viability of the bioelectronic implant in vivo. Insufficient nutrient and oxygen supply is known to be a limiting factor for the survival of encapsulated cells and is a subject of extensive studies. New methods to increase biocompatibility, and to promote vascularization could potentially improve the implant’s engraftment. Although the presence of a mechanical barrier surrounding the therapeutic cells protects them from the host’s immune system, it also limits the diffusion of oxygen, nutrients, and metabolites. In contrast, constructs made out of autologous cells that do not need to be protected from the immune system by a semi-permeable membrane, could be vascularized directly, and benefit from efficient nutrient exchange. As discussed before, such a solution would require personalized engineering of a patient’s own cells and therefore, put a greater emphasis on the safety and immunogenicity of the biological design. Future research in the field of bioelectronic implants will need to align development of the electrogenetic interface with optimization of the implant construction.

Rewiring of endogenous signalling pathways to genomic targets for therapeutic cell reprogramming

Each cell functions as a biological computer that can sense and process signals. Immune cells, specialized in defending organism from pathogens, make a perfect example. For instance, the activation of a T-cell receptor (TCR) activates the killing program of the T-cell. Adding a new input signal by introducing a chimeric antigen receptor allows redirecting that attack to a target of choice. Although such modification alone might be sufficient to destroy some types of cancer cells, other types remain a challenge. Because T-cells integrate multiple signals, which determine their fate, insufficient activation might lead to anergy, while excessive or unspecific activations might cause a life threatening cytokine release syndrome36,260,261. New generations of CAR-Ts are being designed to bypass that limitation. For instance, fourth generation of CAR-Ts, named as TRUCKs, introduced a transgenic immunostimulatory cytokine expression, to improve solid tumor infiltration and killing262. Other designs incorporate logic gates263, or inactivate response to immunosuppressive cytokines264.

A common trend in therapeutic cell engineering focused on creating input-specific programs by engineering artificial receptors215. Some of them were designed to activate

Page 173 of 200

endogenous gene expression via dCas9. Although they perform robustly for some input molecules, others remain a challenge. Two designs that were developed to upregulate expression of endogenous immunostimulatory IL-2 as a response to tumor vasculature- associated VEGF in immune cell lines201,202,265. Only one of them was reported to exhibit 2-3 fold induction upon stimulation with soluble VEGF, however, that effect could not be reproduced in this work202. Meanwhile, GEARs - that were developed in the context of this dissertation - are pathway-specific and can cooperate with natural receptors while optimized for generations in the process of evolution. Although such a solution is not suitable whenever selective activation is required, it can integrate signals from multiple inputs and use it to tune the response. Furthermore, GEARs enable activation of more than one target genes. Although the current toolbox cannot redirect different GEARs to separate targets, extending it with different CRISPR-Cas subtypes that operate with different types of gRNA266, as well as different protein-binding RNA aptamers267 would open a possibility to redirect different signaling pathways independently.

The discovery of CRISPR/Cas9 followed by the invention of catalytically dead Cas9 (dCas9) greatly simplified modulation of endogenous gene expression. As redirecting dCas9 to different targets does not require protein engineering, it quickly gained popularity in synthetic biology and a need for its regulated activation emerged. Inducible expression of the dCas9 protein has slow ON and OFF kinetics due to the transcriptional delay and high stability of dCas9 mRNA and protein. GEARs avoid that limitation because they can be pre-produced and respond immediately upon activation of the signaling pathway.

A key challenge, but also a potential strength is related to context dependence of dCas9- based transcriptional regulation. Introducing a new factor into a chromatin landscape can interfere with natural transcriptional machinery. Hence, calibrating this system to clinical applications could require extensive screening of a target region, as well as extending the choice of effector domains with different transactivation domains268, or epigenetic modulators269. On the other hand, one could attempt to eliminate undesired effects by targeting binding regions of inhibitory transcription regulators, or by designing inhibitory GEARs (iGEARs). Overall, modular GEARs could become the tool of choice for armoring future generations of CARs and TRUCKs.

Page 174 of 200

References

1 Vigne, J. D. The origins of animal domestication and husbandry: a major change in the history of humanity and the biosphere. C R Biol 334, 171-181, doi:10.1016/j.crvi.2010.12.009 (2011). 2 Kafarski, P. Rainbow code of biotechnology. Chemik 66, 811-816 (2012). 3 Roberts, M. A., Cranenburgh, R. M., Stevens, M. P. & Oyston, P. C. Synthetic biology: biology by design. Microbiology 159, 1219-1220, doi:10.1099/mic.0.069724-0 (2013). 4 in Encyclopædia Britannica (ed The Editors of Encyclopaedia Britannica) (Encyclopædia Britannica, inc., 2016). 5 Saltepe, B., Kehribar, E. S., Su Yirmibesoglu, S. S. & Safak Seker, U. O. Cellular Biosensors with Engineered Genetic Circuits. ACS Sens 3, 13-26, doi:10.1021/acssensors.7b00728 (2018). 6 Zhang, W. & Nielsen, D. R. Synthetic biology applications in industrial microbiology. Front Microbiol 5, 451, doi:10.3389/fmicb.2014.00451 (2014). 7 Fesenko, E. & Edwards, R. Plant synthetic biology: a new platform for industrial biotechnology. J Exp Bot 65, 1927-1937, doi:10.1093/jxb/eru070 (2014). 8 Haellman, V. & Fussenegger, M. Synthetic Biology--Toward Therapeutic Solutions. J Mol Biol 428, 945-962, doi:10.1016/j.jmb.2015.08.020 (2016). 9 Kojima, R., Aubel, D. & Fussenegger, M. Toward a world of theranostic medication: Programming biological sentinel systems for therapeutic intervention. Adv Drug Deliv Rev 105, 66-76, doi:10.1016/j.addr.2016.05.006 (2016). 10 Auslander, S. & Fussenegger, M. From gene switches to mammalian designer cells: present and future prospects. Trends Biotechnol 31, 155-168, doi:10.1016/j.tibtech.2012.11.006 (2013). 11 Schukur, L. & Fussenegger, M. Engineering of synthetic gene circuits for (re-)balancing physiological processes in chronic diseases. Wiley interdisciplinary reviews. Systems biology and medicine 8, 402-422, doi:10.1002/wsbm.1345 (2016). 12 Saxena, P. et al. A programmable synthetic lineage-control network that differentiates human IPSCs into glucose-sensitive insulin-secreting beta-like cells. Nature Communications 7, 11247, doi:10.1038/ncomms11247 (2016). 13 Saxena, P., Bojar, D., Zulewski, H. & Fussenegger, M. Generation of glucose-sensitive insulin-secreting beta-like cells from human embryonic stem cells by incorporating a

Page 175 of 200

synthetic lineage-control network. J Biotechnol 259, 39-45, doi:10.1016/j.jbiotec.2017.07.018 (2017). 14 Bashor, C. J. & Collins, J. J. Understanding Biological Regulation Through Synthetic Biology. Annu Rev Biophys 47, 399-423, doi:10.1146/annurev-biophys-070816-033903 (2018). 15 US Food and Drug Administration. Approved Cellular and Gene Therapy Products, [https://www.fda.gov/] 16 Shao, J. et al. Smartphone-controlled optogenetically engineered cells enable semiautomatic glucose homeostasis in diabetic mice. Science translational medicine 9, doi:10.1126/scitranslmed.aal2298 (2017). 17 Folcher, M. et al. Mind-controlled transgene expression by a wireless-powered optogenetic designer cell implant. Nat Commun 5, 5392, doi:10.1038/ncomms6392 (2014). 18 Starzl, T. E. History of clinical transplantation. World J Surg 24, 759-782 (2000). 19 Fischbach, M. A., Bluestone, J. A. & Lim, W. A. Cell-based therapeutics: the next pillar of medicine. Science translational medicine 5, 179ps177, doi:10.1126/scitranslmed.3005568 (2013). 20 Grupp, S. A. & June, C. H. Adoptive cellular therapy. Curr Top Microbiol Immunol 344, 149-172, doi:10.1007/82_2010_94 (2011). 21 Arellano, B., Graber, D. J. & Sentman, C. L. Regulatory T cell-based therapies for autoimmunity. Discov Med 22, 73-80 (2016). 22 Restifo, N. P., Dudley, M. E. & Rosenberg, S. A. Adoptive immunotherapy for cancer: harnessing the T cell response. Nat Rev Immunol 12, 269-281, doi:10.1038/nri3191 (2012). 23 Tao, Z. et al. Cellular immunotherapy of cancer: an overview and future directions. Immunotherapy 9, 589-606, doi:10.2217/imt-2016-0086 (2017). 24 Maus, M. V. et al. Adoptive immunotherapy for cancer or viruses. Annu Rev Immunol 32, 189-225, doi:10.1146/annurev-immunol-032713-120136 (2014). 25 Weiden, P. L. et al. Antileukemic effect of graft-versus-host disease in human recipients of allogeneic-marrow grafts. N Engl J Med 300, 1068-1073, doi:10.1056/NEJM197905103001902 (1979).

Page 176 of 200

26 Rosenberg, S. A. et al. Durable complete responses in heavily pretreated patients with metastatic melanoma using T-cell transfer immunotherapy. Clin Cancer Res 17, 4550- 4557, doi:10.1158/1078-0432.CCR-11-0116 (2011). 27 US National Library of Medicine. Clinical Trials, [https://clinicaltrials.gov/] 28 European Medicines Agency. EU Clinical Trials Register, [https://www.clinicaltrialsregister.eu/] 29 June, C. H., O'Connor, R. S., Kawalekar, O. U., Ghassemi, S. & Milone, M. C. CAR T cell immunotherapy for human cancer. Science 359, 1361-1365, doi:10.1126/science.aar6711 (2018). 30 D’Aloia, M. M., Zizzari, I. G., Sacchetti, B., Pierelli, L. & Alimandi, M. CAR-T cells: the long and winding road to solid tumors. Cell Death & Disease 9, 282, doi:10.1038/s41419-018-0278-6 (2018). 31 Bollard, C. M. et al. Adapting a transforming growth factor beta-related tumor protection strategy to enhance antitumor immunity. Blood 99, 3179-3187 (2002). 32 Kloss, C. C. et al. Dominant-Negative TGF-beta Receptor Enhances PSMA-Targeted Human CAR T Cell Proliferation And Augments Prostate Cancer Eradication. Mol Ther 26, 1855-1866, doi:10.1016/j.ymthe.2018.05.003 (2018). 33 Bendle, G. M., Linnemann, C., Bies, L., Song, J. Y. & Schumacher, T. N. Blockade of TGF-beta signaling greatly enhances the efficacy of TCR gene therapy of cancer. J Immunol 191, 3232-3239, doi:10.4049/jimmunol.1301270 (2013). 34 Mohammed, S. et al. Improving Chimeric Antigen Receptor-Modified T Cell Function by Reversing the Immunosuppressive Tumor Microenvironment of Pancreatic Cancer. Mol Ther 25, 249-258, doi:10.1016/j.ymthe.2016.10.016 (2017). 35 Wang, W. et al. Specificity redirection by CAR with human VEGFR-1 affinity endows T lymphocytes with tumor-killing ability and anti-angiogenic potency. Gene Ther 20, 970-978, doi:10.1038/gt.2013.19 (2013). 36 Shimabukuro-Vornhagen, A. et al. Cytokine release syndrome. J Immunother Cancer 6, 56, doi:10.1186/s40425-018-0343-9 (2018). 37 Chen, K. H. et al. A compound chimeric antigen receptor strategy for targeting multiple myeloma. Leukemia 32, 402-412, doi:10.1038/leu.2017.302 (2018). 38 Grada, Z. et al. TanCAR: A Novel Bispecific Chimeric Antigen Receptor for Cancer Immunotherapy. Mol Ther Nucleic Acids 2, e105, doi:10.1038/mtna.2013.32 (2013).

Page 177 of 200

39 Martyniszyn, A., Krahl, A. C., Andre, M. C., Hombach, A. A. & Abken, H. CD20- CD19 Bispecific CAR T Cells for the Treatment of B-Cell Malignancies. Hum Gene Ther 28, 1147-1157, doi:10.1089/hum.2017.126 (2017). 40 Fedorov, V. D., Themeli, M. & Sadelain, M. PD-1- and CTLA-4-based inhibitory chimeric antigen receptors (iCARs) divert off-target immunotherapy responses. Science translational medicine 5, 215ra172, doi:10.1126/scitranslmed.3006597 (2013). 41 Swanlund, S. L., Scherck, K. A., Metcalfe, S. A. & Jesek-Hale, S. R. Keys to successful self-management of medications. Nurs Sci Q 21, 238-246, doi:10.1177/0894318408319276 (2008). 42 Brown, M. T. & Bussell, J. K. Medication adherence: WHO cares? Mayo Clin Proc 86, 304-314, doi:10.4065/mcp.2010.0575 (2011). 43 Lemstra, M., Nwankwo, C., Bird, Y. & Moraros, J. Primary nonadherence to chronic disease medications: a meta-analysis. Patient Prefer Adherence 12, 721-731, doi:10.2147/PPA.S161151 (2018). 44 Schukur, L., Geering, B., Charpin-El Hamri, G. & Fussenegger, M. Implantable synthetic cytokine converter cells with AND-gate logic treat experimental psoriasis. Science translational medicine 7, 318ra201, doi:10.1126/scitranslmed.aac4964 (2015). 45 Tastanova, A. et al. Synthetic biology-based cellular biomedical tattoo for detection of hypercalcemia associated with cancer. Science translational medicine 10, doi:10.1126/scitranslmed.aap8562 (2018). 46 Xie, M. et al. beta-cell-mimetic designer cells provide closed-loop glycemic control. Science 354, 1296-1301, doi:10.1126/science.aaf4006 (2016). 47 Kushibiki, T., Okawa, S., Hirasawa, T. & Ishihara, M. Optogenetic control of insulin secretion by pancreatic beta-cells in vitro and in vivo. Gene Ther 22, 553-559, doi:10.1038/gt.2015.23 (2015). 48 Stanley, S. A. et al. Radio-wave heating of iron oxide nanoparticles can regulate plasma glucose in mice. Science 336, 604-608, doi:10.1126/science.1216753 (2012). 49 Stanley, S. A., Sauer, J., Kane, R. S., Dordick, J. S. & Friedman, J. M. Remote regulation of glucose homeostasis in mice using genetically encoded nanoparticles. Nature medicine 21, 92-98, doi:10.1038/nm.3730 (2015). 50 Kerr, D., Wizemann, E., Senstius, J., Zacho, M. & Ampudia-Blasco, F. J. Stability and performance of rapid-acting insulin analogs used for continuous subcutaneous insulin

Page 178 of 200

infusion: a systematic review. J Diabetes Sci Technol 7, 1595-1606, doi:10.1177/193229681300700620 (2013). 51 Spitz, F. & Furlong, E. E. Transcription factors: from enhancer binding to developmental control. Nat Rev Genet 13, 613-626, doi:10.1038/nrg3207 (2012). 52 Gossen, M. & Bujard, H. Tight control of gene expression in mammalian cells by tetracycline-responsive promoters. Proc Natl Acad Sci U S A 89, 5547-5551, doi:10.1073/pnas.89.12.5547 (1992). 53 Gossen, M. et al. Transcriptional activation by tetracyclines in mammalian cells. Science 268, 1766-1769 (1995). 54 Saxena, P., Charpin-El Hamri, G., Folcher, M., Zulewski, H. & Fussenegger, M. Synthetic gene network restoring endogenous pituitary-thyroid feedback control in experimental Graves' disease. Proc Natl Acad Sci U S A 113, 1244-1249, doi:10.1073/pnas.1514383113 (2016). 55 Weber, W. & Fussenegger, M. Engineering of synthetic mammalian gene networks. Chem Biol 16, 287-297, doi:10.1016/j.chembiol.2009.02.005 (2009). 56 Rossger, K., Charpin-El-Hamri, G. & Fussenegger, M. A closed-loop synthetic gene circuit for the treatment of diet-induced obesity in mice. Nat Commun 4, 2825, doi:10.1038/ncomms3825 (2013). 57 Ye, H., Daoud-El Baba, M., Peng, R. W. & Fussenegger, M. A synthetic optogenetic transcription device enhances blood-glucose homeostasis in mice. Science 332, 1565- 1568, doi:10.1126/science.1203535 (2011). 58 Ye, H. et al. Pharmaceutically controlled designer circuit for the treatment of the metabolic syndrome. Proc Natl Acad Sci U S A 110, 141-146, doi:10.1073/pnas.1216801110 (2013). 59 Kim, T., Folcher, M., Charpin-El Hamri, G. & Fussenegger, M. A synthetic cGMP- sensitive gene switch providing Viagra((R))-controlled gene expression in mammalian cells and mice. Metabolic engineering 29, 169-179, doi:10.1016/j.ymben.2015.03.016 (2015). 60 Le, X. et al. Heat shock-inducible Cre/Lox approaches to induce diverse types of tumors and hyperplasia in transgenic zebrafish. Proc Natl Acad Sci U S A 104, 9410-9415, doi:10.1073/pnas.0611302104 (2007). 61 Boettcher, M. & McManus, M. T. Choosing the Right Tool for the Job: RNAi, TALEN, or CRISPR. Mol Cell 58, 575-585, doi:10.1016/j.molcel.2015.04.028 (2015).

Page 179 of 200

62 Maji, B. et al. Multidimensional chemical control of CRISPR-Cas9. Nat Chem Biol 13, 9-11, doi:10.1038/nchembio.2224 (2017). 63 Lu, J. et al. Multimode drug inducible CRISPR/Cas9 devices for transcriptional activation and genome editing. Nucleic Acids Res 46, e25, doi:10.1093/nar/gkx1222 (2018). 64 Gao, Y. et al. Complex transcriptional modulation with orthogonal and inducible dCas9 regulators. Nat Methods 13, 1043-1049, doi:10.1038/nmeth.4042 (2016). 65 Polstein, L. R. & Gersbach, C. A. A light-inducible CRISPR-Cas9 system for control of endogenous gene activation. Nat Chem Biol 11, 198-200, doi:10.1038/nchembio.1753 (2015). 66 Zhou, X. X. et al. A Single-Chain Photoswitchable CRISPR-Cas9 Architecture for Light-Inducible Gene Editing and Transcription. ACS Chem Biol 13, 443-448, doi:10.1021/acschembio.7b00603 (2018). 67 Saito, H. et al. Synthetic translational regulation by an L7Ae-kink-turn RNP switch. Nat Chem Biol 6, 71-78, doi:10.1038/nchembio.273 (2010). 68 An, C. I., Trinh, V. B. & Yokobayashi, Y. Artificial control of gene expression in mammalian cells by modulating RNA interference through aptamer-small molecule interaction. RNA 12, 710-716, doi:10.1261/rna.2299306 (2006). 69 Beisel, C. L., Chen, Y. Y., Culler, S. J., Hoff, K. G. & Smolke, C. D. Design of small molecule-responsive microRNAs based on structural requirements for Drosha processing. Nucleic Acids Res 39, 2981-2994, doi:10.1093/nar/gkq954 (2011). 70 Auslander, S., Ketzer, P. & Hartig, J. S. A ligand-dependent hammerhead ribozyme switch for controlling mammalian gene expression. Mol Biosyst 6, 807-814, doi:10.1039/b923076a (2010). 71 Weigand, J. E. & Suess, B. Tetracycline aptamer-controlled regulation of pre-mRNA splicing in yeast. Nucleic Acids Res 35, 4179-4185, doi:10.1093/nar/gkm425 (2007). 72 Bojar, D., Fuhrer, T. & Fussenegger, M. Purity by design: Reducing impurities in bioproduction by stimulus-controlled global translational downregulation of non- product proteins. Metabolic engineering 52, 110-123, doi:10.1016/j.ymben.2018.11.007 (2019). 73 Chassin, H. et al. A modular degron library for synthetic circuits in mammalian cells. Nat Commun 10, 2013, doi:10.1038/s41467-019-09974-5 (2019).

Page 180 of 200

74 Nishimura, K., Fukagawa, T., Takisawa, H., Kakimoto, T. & Kanemaki, M. An auxin- based degron system for the rapid depletion of proteins in nonplant cells. Nat Methods 6, 917-922, doi:10.1038/nmeth.1401 (2009). 75 Neklesa, T. K. et al. Small-molecule hydrophobic tagging-induced degradation of HaloTag fusion proteins. Nat Chem Biol 7, 538-543, doi:10.1038/nchembio.597 (2011). 76 Banaszynski, L. A., Chen, L. C., Maynard-Smith, L. A., Ooi, A. G. & Wandless, T. J. A rapid, reversible, and tunable method to regulate protein function in living cells using synthetic small molecules. Cell 126, 995-1004, doi:10.1016/j.cell.2006.07.025 (2006). 77 Bonger, K. M., Chen, L. C., Liu, C. W. & Wandless, T. J. Small-molecule displacement of a cryptic degron causes conditional protein degradation. Nat Chem Biol 7, 531-537, doi:10.1038/nchembio.598 (2011). 78 Iwamoto, M., Bjorklund, T., Lundberg, C., Kirik, D. & Wandless, T. J. A general chemical method to regulate protein stability in the mammalian central nervous system. Chem Biol 17, 981-988, doi:10.1016/j.chembiol.2010.07.009 (2010). 79 Miyazaki, Y., Imoto, H., Chen, L. C. & Wandless, T. J. Destabilizing domains derived from the human estrogen receptor. J Am Chem Soc 134, 3942-3945, doi:10.1021/ja209933r (2012). 80 Schneekloth, J. S., Jr. et al. Chemical genetic control of protein levels: selective in vivo targeted degradation. J Am Chem Soc 126, 3748-3754, doi:10.1021/ja039025z (2004). 81 Rivera, V. M. et al. Regulation of protein secretion through controlled aggregation in the endoplasmic reticulum. Science 287, 826-830 (2000). 82 Yissachar, N. et al. Dynamic response diversity of NFAT isoforms in individual living cells. Mol Cell 49, 322-330, doi:10.1016/j.molcel.2012.11.003 (2013). 83 Shamir, M., Bar-On, Y., Phillips, R. & Milo, R. SnapShot: Timescales in Cell Biology. Cell 164, 1302-1302 e1301, doi:10.1016/j.cell.2016.02.058 (2016). 84 Audibert, A., Weil, D. & Dautry, F. In vivo kinetics of mRNA splicing and transport in mammalian cells. Mol Cell Biol 22, 6706-6718 (2002). 85 Thor, F., Gautschi, M., Geiger, R. & Helenius, A. Bulk flow revisited: transport of a soluble protein in the secretory pathway. Traffic 10, 1819-1830, doi:10.1111/j.1600- 0854.2009.00989.x (2009). 86 Kasai, H., Suzuki, T., Liu, T. T., Kishimoto, T. & Takahashi, N. Fast and cAMP- sensitive mode of Ca(2+)-dependent exocytosis in pancreatic beta-cells. Diabetes 51 Suppl 1, S19-24 (2002).

Page 181 of 200

87 Henquin, J. C., Dufrane, D., Kerr-Conte, J. & Nenquin, M. Dynamics of glucose- induced insulin secretion in normal human islets. Am J Physiol Endocrinol Metab 309, E640-650, doi:10.1152/ajpendo.00251.2015 (2015). 88 Schweinfest, C. W., Jorcyk, C. L., Fujiwara, S. & Papas, T. S. A heat-shock-inducible eukaryotic expression vector. Gene 71, 207-210 (1988). 89 Andersson, H. A., Kim, Y. S., O'Neill, B. E., Shi, Z. Z. & Serda, R. E. HSP70 promoter- driven activation of gene expression for immunotherapy using gold nanorods and near infrared light. Vaccines 2, 216-227, doi:10.3390/vaccines2020216 (2014). 90 Ortner, V. et al. Magnetic field-controlled gene expression in encapsulated cells. Journal of controlled release : official journal of the Controlled Release Society 158, 424-432, doi:10.1016/j.jconrel.2011.12.006 (2012). 91 Huang, H., Delikanli, S., Zeng, H., Ferkey, D. M. & Pralle, A. Remote control of ion channels and neurons through magnetic-field heating of nanoparticles. Nature nanotechnology 5, 602-606, doi:10.1038/nnano.2010.125 (2010). 92 Rost, B. R., Schneider-Warme, F., Schmitz, D. & Hegemann, P. Optogenetic Tools for Subcellular Applications in Neuroscience. Neuron 96, 572-603, doi:10.1016/j.neuron.2017.09.047 (2017). 93 Kim, T., Folcher, M., Doaud-El Baba, M. & Fussenegger, M. A synthetic erectile optogenetic stimulator enabling blue-light-inducible penile erection. Angewandte Chemie 54, 5933-5938, doi:10.1002/anie.201412204 (2015). 94 Pudasaini, A., El-Arab, K. K. & Zoltowski, B. D. LOV-based optogenetic devices: light- driven modules to impart photoregulated control of cellular signaling. Front Mol Biosci 2, 18, doi:10.3389/fmolb.2015.00018 (2015). 95 Weber, W. et al. A synthetic mammalian electro-genetic transcription circuit. Nucleic Acids Res 37, e33, doi:10.1093/nar/gkp014 (2009). 96 Do, M. T. et al. Photon capture and signalling by melanopsin retinal ganglion cells. Nature 457, 281-287, doi:10.1038/nature07682 (2009). 97 Sharrock, R. A. The phytochrome red/far-red photoreceptor superfamily. Genome Biol 9, 230, doi:10.1186/gb-2008-9-8-230 (2008). 98 Tschirhart, T. et al. Electronic control of gene expression and cell behaviour in Escherichia coli through redox signalling. Nat Commun 8, 14030, doi:10.1038/ncomms14030 (2017).

Page 182 of 200

99 Wang, X., Chen, X. & Yang, Y. Spatiotemporal control of gene expression by a light- switchable transgene system. Nat Methods 9, 266-269, doi:10.1038/nmeth.1892 (2012). 100 Meister, M. Physical limits to magnetogenetics. eLife 5, doi:10.7554/eLife.17210 (2016). 101 Anikeeva, P. & Jasanoff, A. Problems on the back of an envelope. eLife 5, doi:10.7554/eLife.19569 (2016). 102 Ito, A., Shinkai, M., Honda, H. & Kobayashi, T. Heat-inducible TNF-alpha gene therapy combined with hyperthermia using magnetic nanoparticles as a novel tumor- targeted therapy. Cancer Gene Ther 8, 649-654, doi:10.1038/sj.cgt.7700357 (2001). 103 Wheeler, M. A. et al. Genetically targeted magnetic control of the nervous system. Nat Neurosci 19, 756-761, doi:10.1038/nn.4265 (2016). 104 Muller, K. et al. A red/far-red light-responsive bi-stable toggle switch to control gene expression in mammalian cells. Nucleic Acids Res 41, e77, doi:10.1093/nar/gkt002 (2013). 105 Duncan, M. D. & Wilkes, D. S. Transplant-related immunosuppression: a review of immunosuppression and pulmonary infections. Proc Am Thorac Soc 2, 449-455, doi:10.1513/pats.200507-073JS (2005). 106 Hwa, A. J. & Weir, G. C. Transplantation of Macroencapsulated Insulin-Producing Cells. Curr Diab Rep 18, 50, doi:10.1007/s11892-018-1028-y (2018). 107 Desai, T. A., Hansford, D. J. & Ferrari, M. Micromachined interfaces: new approaches in cell immunoisolation and biomolecular separation. Biomol Eng 17, 23-36 (2000). 108 Lanza, R. P., Hayes, J. L. & Chick, W. L. Encapsulated cell technology. Nat Biotechnol 14, 1107-1111, doi:10.1038/nbt0996-1107 (1996). 109 Murua, A. et al. Cell microencapsulation technology: towards clinical application. Journal of controlled release : official journal of the Controlled Release Society 132, 76-83, doi:10.1016/j.jconrel.2008.08.010 (2008). 110 Olabisi, R. M. Cell microencapsulation with synthetic polymers. J Biomed Mater Res A 103, 846-859, doi:10.1002/jbm.a.35205 (2015). 111 Orive, G. et al. Cell encapsulation: technical and clinical advances. Trends Pharmacol Sci 36, 537-546, doi:10.1016/j.tips.2015.05.003 (2015). 112 Algire, G. H. & Legallais, F. Y. Recent developments in the transparent-chamber technique as adapted to the mouse. J Natl Cancer Inst 10, 225-253, incl 228 pl (1949).

Page 183 of 200

113 Algire, G. H. & Legallais, F. Y. The transparent chamber technique in the mouse in the study of tumor histo-physiology. Fed Proc 7, 268 (1948). 114 Brauker, J. H. et al. Neovascularization of synthetic membranes directed by membrane microarchitecture. J Biomed Mater Res 29, 1517-1524, doi:10.1002/jbm.820291208 (1995). 115 Lathuiliere, A., Cosson, S., Lutolf, M. P., Schneider, B. L. & Aebischer, P. A high- capacity cell macroencapsulation system supporting the long-term survival of genetically engineered allogeneic cells. Biomaterials 35, 779-791, doi:10.1016/j.biomaterials.2013.09.071 (2014). 116 Strautz, R. L. Studies of hereditary-obese mice (obob) after implantation of pancreatic islets in Millipore filter capsules. Diabetologia 6, 306-312 (1970). 117 Scharp, D. W. et al. Protection of encapsulated human islets implanted without immunosuppression in patients with type I or type II diabetes and in nondiabetic control subjects. Diabetes 43, 1167-1170 (1994). 118 Altman, J. J. et al. Encapsulated human islet transplants in diabetic rats. Trans Am Soc Artif Intern Organs 30, 382-386 (1984). 119 Lacy, P. E., Hegre, O. D., Gerasimidi-Vazeou, A., Gentile, F. T. & Dionne, K. E. Maintenance of normoglycemia in diabetic mice by subcutaneous xenografts of encapsulated islets. Science 254, 1782-1784 (1991). 120 Klomp, G. F., Ronel, S. H., Hashiguchi, H., D'Andrea, M. & Dobelle, W. H. Hydrogels for encapsulation of pancreatic islet cells. Trans Am Soc Artif Intern Organs 25, 74-76 (1979). 121 Storrs, R., Dorian, R., King, S. R., Lakey, J. & Rilo, H. Preclinical development of the Islet Sheet. Ann N Y Acad Sci 944, 252-266 (2001). 122 Evron, Y. et al. Long-term viability and function of transplanted islets macroencapsulated at high density are achieved by enhanced oxygen supply. Sci Rep 8, 6508, doi:10.1038/s41598-018-23862-w (2018). 123 Sun, A. M., Parisius, W., Healy, G. M., Vacek, I. & Macmorine, H. G. The use, in diabetic rats and monkeys, of artificial capillary units containing cultured islets of Langerhans (artificial endocrine pancreas). Diabetes 26, 1136-1139 (1977). 124 Scharp, D. W., Mason, N. S. & Sparks, R. E. Islet immuno-isolation: the use of hybrid artificial organs to prevent islet tissue rejection. World J Surg 8, 221-229 (1984).

Page 184 of 200

125 Ikeda, H. et al. A newly developed bioartificial pancreas successfully controls blood glucose in totally pancreatectomized diabetic pigs. Tissue Eng 12, 1799-1809, doi:10.1089/ten.2006.12.1799 (2006). 126 Song, S. et al. An intravascular bioartificial pancreas device (iBAP) with silicon nanopore membranes (SNM) for islet encapsulation under convective mass transport. Lab Chip 17, 1778-1792, doi:10.1039/c7lc00096k (2017). 127 Prochorov, A. V., Tretjak, S. I., Goranov, V. A., Glinnik, A. A. & Goltsev, M. V. Treatment of insulin dependent diabetes mellitus with intravascular transplantation of pancreatic islet cells without immunosuppressive therapy. Adv Med Sci 53, 240-244 (2008). 128 Vaithilingam, V. & Tuch, B. E. Islet transplantation and encapsulation: an update on recent developments. Rev Diabet Stud 8, 51-67, doi:10.1900/RDS.2011.8.51 (2011). 129 Krishnan, R., Alexander, M., Robles, L., Foster, C. E., 3rd & Lakey, J. R. Islet and stem cell encapsulation for clinical transplantation. Rev Diabet Stud 11, 84-101, doi:10.1900/RDS.2014.11.84 (2014). 130 Vos, P. d., Andersson, A., Tam, S. K., Faas, M. M. & Halle, J. P. Advances and Barriers in Mammalian Cell Encapsulation for Treatment of Diabetes. Immunology, Endocrine & Metabolic Agents - Medicinal ChemistryCurrent Medicinal Chemistry - Immunology, Endocrine & Metabolic Agents) 6, 139-153, doi:10.2174/187152206776359948 (2006). 131 Katsarou, A. et al. Type 1 diabetes mellitus. Nature Reviews Disease Primers 3, 17016, doi:10.1038/nrdp.2017.16 (2017). 132 Shapiro, A. M. Islet transplantation in type 1 diabetes: ongoing challenges, refined procedures, and long-term outcome. Rev Diabet Stud 9, 385-406, doi:10.1900/RDS.2012.9.385 (2012). 133 Shapiro, A. M. et al. Strategic opportunities in clinical islet transplantation. Transplantation 79, 1304-1307 (2005). 134 Skelin, M., Rupnik, M. & Cencic, A. Pancreatic beta cell lines and their applications in diabetes mellitus research. Altex 27, 105-113 (2010). 135 Soon-Shiong, P. et al. Insulin independence in a type 1 diabetic patient after encapsulated islet transplantation. Lancet 343, 950-951 (1994). 136 Qi, M. Transplantation of Encapsulated Pancreatic Islets as a Treatment for Patients with Type 1 Diabetes Mellitus. Adv Med 2014, 429710, doi:10.1155/2014/429710 (2014).

Page 185 of 200

137 Aye, M. M. & Atkin, S. L. Patient safety and minimizing risk with insulin administration - role of insulin degludec. Drug, healthcare and patient safety 6, 55-67, doi:10.2147/DHPS.S59566 (2014). 138 Goto, A., Arah, O. A., Goto, M., Terauchi, Y. & Noda, M. Severe hypoglycaemia and cardiovascular disease: systematic review and meta-analysis with bias analysis. Bmj 347, f4533, doi:10.1136/bmj.f4533 (2013). 139 Leese, G. P. et al. Frequency of severe hypoglycemia requiring emergency treatment in type 1 and type 2 diabetes: a population-based study of health service resource use. Diabetes care 26, 1176-1180 (2003). 140 Uduku, C. & Oliver, N. Pharmacological aspects of closed loop insulin delivery for type 1 diabetes. Current opinion in pharmacology 36, 29-33, doi:10.1016/j.coph.2017.07.006 (2017). 141 Fussenegger, M. et al. Streptogramin-based gene regulation systems for mammalian cells. Nat Biotechnol 18, 1203-1208, doi:10.1038/81208 (2000). 142 Weber, W. et al. Macrolide-based transgene control in mammalian cells and mice. Nat Biotechnol 20, 901-907, doi:10.1038/nbt731 (2002). 143 Auner, B. G., Valenta, C. & Hadgraft, J. Influence of phloretin and 6-ketocholestanol on the skin permeation of sodium-fluorescein. Journal of controlled release : official journal of the Controlled Release Society 89, 321-328 (2003). 144 Gitzinger, M., Kemmer, C., El-Baba, M. D., Weber, W. & Fussenegger, M. Controlling transgene expression in subcutaneous implants using a skin lotion containing the apple metabolite phloretin. Proc Natl Acad Sci U S A 106, 10638-10643, doi:10.1073/pnas.0901501106 (2009). 145 Gitzinger, M. et al. The food additive vanillic acid controls transgene expression in mammalian cells and mice. Nucleic Acids Res 40, e37, doi:10.1093/nar/gkr1251 (2012). 146 Wang, H., Ye, H., Xie, M., Daoud El-Baba, M. & Fussenegger, M. Cosmetics-triggered percutaneous remote control of transgene expression in mice. Nucleic Acids Res 43, e91, doi:10.1093/nar/gkv326 (2015). 147 Prindle, A. et al. Rapid and tunable post-translational coupling of genetic circuits. Nature 508, 387-391, doi:10.1038/nature13238 (2014). 148 Weber, W. et al. Gas-inducible transgene expression in mammalian cells and mice. Nat Biotechnol 22, 1440-1444, doi:10.1038/nbt1021 (2004).

Page 186 of 200

149 Katz, D. & Akiyama, T. Pacemaker longevity: the world's longest-lasting VVI pacemaker. Ann Noninvasive Electrocardiol 12, 223-226, doi:10.1111/j.1542- 474X.2007.00165.x (2007). 150 Stockley, J. H. et al. Surpassing light-induced cell damage in vitro with novel cell culture media. Sci Rep 7, 849, doi:10.1038/s41598-017-00829-x (2017). 151 Hinds, T. D., Jr. et al. Biliverdin Reductase A Attenuates Hepatic Steatosis by Inhibition of Glycogen Synthase Kinase (GSK) 3beta Phosphorylation of Serine 73 of Peroxisome Proliferator-activated Receptor (PPAR) alpha. The Journal of biological chemistry 291, 25179-25191, doi:10.1074/jbc.M116.731703 (2016). 152 Uda, Y. et al. Efficient synthesis of phycocyanobilin in mammalian cells for optogenetic control of cell signaling. Proc Natl Acad Sci U S A 114, 11962-11967, doi:10.1073/pnas.1707190114 (2017). 153 Hu, Z. et al. Biliverdin Reductase A (BVRA) Mediates Macrophage Expression of Interleukin-10 in Injured Kidney. Int J Mol Sci 16, 22621-22635, doi:10.3390/ijms160922621 (2015). 154 Beyer, H. M. et al. Red Light-Regulated Reversible Nuclear Localization of Proteins in Mammalian Cells and Zebrafish. ACS Synth Biol 4, 951-958, doi:10.1021/acssynbio.5b00004 (2015). 155 Jarockyte, G. et al. Accumulation and Toxicity of Superparamagnetic Iron Oxide Nanoparticles in Cells and Experimental Animals. Int J Mol Sci 17, doi:10.3390/ijms17081193 (2016). 156 Maher, B. A. et al. Magnetite pollution nanoparticles in the human brain. Proc Natl Acad Sci U S A 113, 10797-10801, doi:10.1073/pnas.1605941113 (2016). 157 Ashcroft, F. M. & Rorsman, P. K(ATP) channels and islet hormone secretion: new insights and controversies. Nature reviews. Endocrinology 9, 660-669, doi:10.1038/nrendo.2013.166 (2013). 158 Buzzetti, R., Zampetti, S. & Maddaloni, E. Adult-onset autoimmune diabetes: current knowledge and implications for management. Nature reviews. Endocrinology 13, 674- 686, doi:10.1038/nrendo.2017.99 (2017). 159 Jacobs-Tulleneers-Thevissen, D. et al. Sustained function of alginate-encapsulated human islet cell implants in the peritoneal cavity of mice leading to a pilot study in a type 1 diabetic patient. Diabetologia 56, 1605-1614, doi:10.1007/s00125-013-2906-0 (2013).

Page 187 of 200

160 Desai, T. & Shea, L. D. Advances in islet encapsulation technologies. Nature reviews. Drug discovery 16, 338-350, doi:10.1038/nrd.2016.232 (2017). 161 Zamponi, G. W., Striessnig, J., Koschak, A. & Dolphin, A. C. The Physiology, Pathology, and Pharmacology of Voltage-Gated Calcium Channels and Their Future Therapeutic Potential. Pharmacological reviews 67, 821-870, doi:10.1124/pr.114.009654 (2015). 162 D'Arco, M. & Dolphin, A. C. L-type calcium channels: on the fast track to nuclear signaling. Science signaling 5, pe34, doi:10.1126/scisignal.2003355 (2012). 163 Kim, T. et al. The biochemical activation of T-type Ca2+ channels in HEK293 cells stably expressing alpha1G and Kir2.1 subunits. Biochemical and biophysical research communications 324, 401-408, doi:10.1016/j.bbrc.2004.09.069 (2004). 164 Llucia-Valldeperas, A. et al. Electrical stimulation of cardiac adipose tissue-derived progenitor cells modulates cell phenotype and genetic machinery. Journal of tissue engineering and regenerative medicine 9, E76-83, doi:10.1002/term.1710 (2015). 165 Thelen, M. H., Simonides, W. S. & van Hardeveld, C. Electrical stimulation of C2C12 myotubes induces contractions and represses thyroid-hormone-dependent transcription of the fast-type sarcoplasmic-reticulum Ca2+-ATPase gene. The Biochemical journal 321 ( Pt 3), 845-848 (1997). 166 Manabe, Y. et al. Characterization of an acute muscle contraction model using cultured C2C12 myotubes. PloS one 7, e52592, doi:10.1371/journal.pone.0052592 (2012). 167 McCluskey, J. T. et al. Development and functional characterization of insulin-releasing human pancreatic beta cell lines produced by electrofusion. The Journal of biological chemistry 286, 21982-21992, doi:10.1074/jbc.M111.226795 (2011). 168 Burns, S. M. et al. High-throughput luminescent reporter of insulin secretion for discovering regulators of pancreatic Beta-cell function. Cell metabolism 21, 126-137, doi:10.1016/j.cmet.2014.12.010 (2015). 169 Gromada, J., Chabosseau, P. & Rutter, G. A. The alpha-cell in diabetes mellitus. Nature reviews. Endocrinology 14, 694-704, doi:10.1038/s41574-018-0097-y (2018). 170 Yamashita, H., Nguyen, D. T. & Chung, E. Blood-based assay with secreted Gaussia luciferase to monitor tumor metastasis. Methods Mol Biol 1098, 145-151, doi:10.1007/978-1-62703-718-1_12 (2014). 171 Rattner, B. in Host Response to Biomaterials (ed SF Badylak) Ch. 3, 37-51 (Academic Press, 2015).

Page 188 of 200

172 Russell, S. J. et al. Day and night glycaemic control with a bionic pancreas versus conventional insulin pump therapy in preadolescent children with type 1 diabetes: a randomised crossover trial. Lancet Diabetes Endocrinol 4, 233-243, doi:10.1016/S2213-8587(15)00489-1 (2016). 173 Hollander, P. A. et al. Importance of early insulin secretion: comparison of nateglinide and glyburide in previously diet-treated patients with type 2 diabetes. Diabetes care 24, 983-988 (2001). 174 Kahn, S. E. et al. Importance of early phase insulin secretion to intravenous glucose tolerance in subjects with type 2 diabetes mellitus. J Clin Endocrinol Metab 86, 5824- 5829, doi:10.1210/jcem.86.12.8105 (2001). 175 Mond, H. G. & Freitag, G. The cardiac implantable electronic device power source: evolution and revolution. Pacing and clinical electrophysiology : PACE 37, 1728-1745, doi:10.1111/pace.12526 (2014). 176 Russell, S. J. et al. Outpatient glycemic control with a bionic pancreas in type 1 diabetes. N Engl J Med 371, 313-325, doi:10.1056/NEJMoa1314474 (2014). 177 Schmelzeisen-Redeker, G. et al. Time Delay of CGM Sensors: Relevance, Causes, and Countermeasures. J Diabetes Sci Technol 9, 1006-1015, doi:10.1177/1932296815590154 (2015). 178 Keenan, D. B., Mastrototaro, J. J., Voskanyan, G. & Steil, G. M. Delays in minimally invasive continuous glucose monitoring devices: a review of current technology. J Diabetes Sci Technol 3, 1207-1214, doi:10.1177/193229680900300528 (2009). 179 HENRY, R. R. et al. Initial Clinical Evaluation of VC-01TM Combination Product—A Stem Cell–Derived Islet Replacement for Type 1 Diabetes (T1D). Diabetes 67, 138- OR, doi:10.2337/db18-138-OR (2018). 180 Motulsky, H. J. & Brown, R. E. Detecting outliers when fitting data with nonlinear regression - a new method based on robust nonlinear regression and the false discovery rate. BMC Bioinformatics 7, 123, doi:10.1186/1471-2105-7-123 (2006). 181 Xu, W. & Lipscombe, D. Neuronal Ca(V)1.3alpha(1) L-type channels activate at relatively hyperpolarized membrane potentials and are incompletely inhibited by dihydropyridines. J Neurosci 21, 5944-5951 (2001). 182 Lin, Y., McDonough, S. I. & Lipscombe, D. Alternative splicing in the voltage-sensing region of N-Type CaV2.2 channels modulates channel kinetics. J Neurophysiol 92, 2820-2830, doi:10.1152/jn.00048.2004 (2004).

Page 189 of 200

183 Jiang, Y. Q., Andrade, A. & Lipscombe, D. Spinal morphine but not ziconotide or gabapentin analgesia is affected by alternative splicing of voltage-gated calcium channel CaV2.2 pre-mRNA. Mol Pain 9, 67, doi:10.1186/1744-8069-9-67 (2013). 184 Cameron, K. N., Solis, E., Jr., Ruchala, I., De Felice, L. J. & Eltit, J. M. Amphetamine activates calcium channels through dopamine transporter-mediated depolarization. Cell Calcium 58, 457-466, doi:10.1016/j.ceca.2015.06.013 (2015). 185 Mates, L. et al. Molecular evolution of a novel hyperactive Sleeping Beauty transposase enables robust stable gene transfer in vertebrates. Nat Genet 41, 753-761, doi:10.1038/ng.343 (2009). 186 Yu, C. R. et al. Spontaneous neural activity is required for the establishment and maintenance of the olfactory sensory map. Neuron 42, 553-566, doi:10.1016/s0896- 6273(04)00224-7 (2004). 187 Follenzi, A., Sabatino, G., Lombardo, A., Boccaccio, C. & Naldini, L. Efficient gene delivery and targeted expression to hepatocytes in vivo by improved lentiviral vectors. Hum Gene Ther 13, 243-260, doi:10.1089/10430340252769770 (2002). 188 Konermann, S. et al. Genome-scale transcriptional activation by an engineered CRISPR-Cas9 complex. Nature 517, 583-588, doi:10.1038/nature14136 (2015). 189 Kowarz, E., Loscher, D. & Marschalek, R. Optimized Sleeping Beauty transposons rapidly generate stable transgenic cell lines. Biotechnol J 10, 647-653, doi:10.1002/biot.201400821 (2015). 190 Aichberger, K. J. et al. Low-level expression of proapoptotic Bcl-2-interacting mediator in leukemic cells in patients with chronic myeloid leukemia: role of BCR/ABL, characterization of underlying signaling pathways, and reexpression by novel pharmacologic compounds. Cancer Res 65, 9436-9444, doi:10.1158/0008-5472.CAN- 05-0972 (2005). 191 Chmielewski, M., Hombach, A. A. & Abken, H. Of CARs and TRUCKs: chimeric antigen receptor (CAR) T cells engineered with an inducible cytokine to modulate the tumor stroma. Immunol Rev 257, 83-90, doi:10.1111/imr.12125 (2014). 192 Adachi, K. et al. IL-7 and CCL19 expression in CAR-T cells improves immune cell infiltration and CAR-T cell survival in the tumor. Nat Biotechnol, doi:10.1038/nbt.4086 (2018).

Page 190 of 200

193 Zhang, L. et al. Improving adoptive T cell therapy by targeting and controlling IL-12 expression to the tumor environment. Mol Ther 19, 751-759, doi:10.1038/mt.2010.313 (2011). 194 Chulpanova, D. S. et al. Application of Mesenchymal Stem Cells for Therapeutic Agent Delivery in Anti-tumor Treatment. Front Pharmacol 9, 259, doi:10.3389/fphar.2018.00259 (2018). 195 Levy, O. et al. mRNA-engineered mesenchymal stem cells for targeted delivery of interleukin-10 to sites of inflammation. Blood 122, e23-32, doi:10.1182/blood-2013-04- 495119 (2013). 196 Schwarz, K. A. & Leonard, J. N. Engineering cell-based therapies to interface robustly with host physiology. Adv Drug Deliv Rev 105, 55-65, doi:10.1016/j.addr.2016.05.019 (2016). 197 Davila, M. L. et al. Efficacy and toxicity management of 19-28z CAR T cell therapy in B cell acute lymphoblastic leukemia. Science translational medicine 6, 224ra225, doi:10.1126/scitranslmed.3008226 (2014). 198 Wu, C. Y., Roybal, K. T., Puchner, E. M., Onuffer, J. & Lim, W. A. Remote control of therapeutic T cells through a small molecule-gated chimeric receptor. Science 350, aab4077, doi:10.1126/science.aab4077 (2015). 199 Straathof, K. C. et al. An inducible caspase 9 safety switch for T-cell therapy. Blood 105, 4247-4254, doi:10.1182/blood-2004-11-4564 (2005). 200 Kloss, C. C., Condomines, M., Cartellieri, M., Bachmann, M. & Sadelain, M. Combinatorial antigen recognition with balanced signaling promotes selective tumor eradication by engineered T cells. Nat Biotechnol 31, 71-75, doi:10.1038/nbt.2459 (2013). 201 Baeumler, T. A., Ahmed, A. A. & Fulga, T. A. Engineering Synthetic Signaling Pathways with Programmable dCas9-Based Chimeric Receptors. Cell reports 20, 2639- 2653, doi:10.1016/j.celrep.2017.08.044 (2017). 202 Schwarz, K. A., Daringer, N. M., Dolberg, T. B. & Leonard, J. N. Rewiring human cellular input-output using modular extracellular sensors. Nat Chem Biol 13, 202-209, doi:10.1038/nchembio.2253 (2017). 203 Saxena, P. et al. A programmable synthetic lineage-control network that differentiates human IPSCs into glucose-sensitive insulin-secreting beta-like cells. Nature communications 7, 11247, doi:10.1038/ncomms11247 (2016).

Page 191 of 200

204 Clapham, D. E. Calcium signaling. Cell 131, 1047-1058, doi:10.1016/j.cell.2007.11.028 (2007). 205 Lawrence, M. C., Bhatt, H. S. & Easom, R. A. NFAT regulates insulin gene promoter activity in response to synergistic pathways induced by glucose and glucagon-like peptide-1. Diabetes 51, 691-698 (2002). 206 Shaw, J. P. et al. Identification of a putative regulator of early T cell activation genes. Science 241, 202-205, doi:10.1126/science.3260404 (1988). 207 Vom Berg, J. et al. Intratumoral IL-12 combined with CTLA-4 blockade elicits T cell- mediated glioma rejection. J Exp Med 210, 2803-2811, doi:10.1084/jem.20130678 (2013). 208 Ma, X. TNF-alpha and IL-12: a balancing act in macrophage functioning. Microbes Infect 3, 121-129 (2001). 209 Zakharova, M. & Ziegler, H. K. Paradoxical anti-inflammatory actions of TNF-alpha: inhibition of IL-12 and IL-23 via TNF receptor 1 in macrophages and dendritic cells. J Immunol 175, 5024-5033 (2005). 210 Korc, M. & Friesel, R. E. The role of fibroblast growth factors in tumor growth. Curr Cancer Drug Targets 9, 639-651 (2009). 211 Scheller, L., Strittmatter, T., Fuchs, D., Bojar, D. & Fussenegger, M. Generalized extracellular molecule sensor platform for programming cellular behavior. Nature Chemical Biology, doi:10.1038/s41589-018-0046-z (2018). 212 Folcher, M., Xie, M., Spinnler, A. & Fussenegger, M. Synthetic mammalian trigger- controlled bipartite transcription factors. Nucleic Acids Res 41, e134, doi:10.1093/nar/gkt405 (2013). 213 Muller, K., Engesser, R., Timmer, J., Zurbriggen, M. D. & Weber, W. Orthogonal optogenetic triple-gene control in Mammalian cells. ACS Synth Biol 3, 796-801, doi:10.1021/sb500305v (2014). 214 Liu, Y., Charpin-El Hamri, G., Ye, H. & Fussenegger, M. A synthetic free fatty acid- regulated transgene switch in mammalian cells and mice. Nucleic Acids Res 46, 9864- 9874, doi:10.1093/nar/gky805 (2018). 215 Santorelli, M., Lam, C. & Morsut, L. Synthetic development: building mammalian multicellular structures with artificial genetic programs. Curr Opin Biotechnol 59, 130- 140, doi:10.1016/j.copbio.2019.03.016 (2019).

Page 192 of 200

216 Kipniss, N. H. et al. Engineering cell sensing and responses using a GPCR-coupled CRISPR-Cas system. Nat Commun 8, 2212, doi:10.1038/s41467-017-02075-1 (2017). 217 Mandegar, M. A. et al. CRISPR Interference Efficiently Induces Specific and Reversible Gene Silencing in Human iPSCs. Cell stem cell 18, 541-553, doi:10.1016/j.stem.2016.01.022 (2016). 218 Gilbert, L. A. et al. Genome-Scale CRISPR-Mediated Control of Gene Repression and Activation. Cell 159, 647-661, doi:10.1016/j.cell.2014.09.029 (2014). 219 Aubrey, B. J. et al. An inducible lentiviral guide RNA platform enables the identification of tumor-essential genes and tumor-promoting mutations in vivo. Cell reports 10, 1422-1432, doi:10.1016/j.celrep.2015.02.002 (2015). 220 Balboa, D. et al. Conditionally Stabilized dCas9 Activator for Controlling Gene Expression in Human Cell Reprogramming and Differentiation. Stem Cell Reports 5, 448-459, doi:10.1016/j.stemcr.2015.08.001 (2015). 221 Shariati, S. A. et al. Reversible Disruption of Specific Transcription Factor-DNA Interactions Using CRISPR/Cas9. Mol Cell 74, 622-633 e624, doi:10.1016/j.molcel.2019.04.011 (2019). 222 Kiani, S. et al. Cas9 gRNA engineering for genome editing, activation and repression. Nat Methods 12, 1051-1054, doi:10.1038/nmeth.3580 (2015). 223 Kiani, S. et al. CRISPR transcriptional repression devices and layered circuits in mammalian cells. Nat Methods 11, 723-726, doi:10.1038/nmeth.2969 (2014). 224 Xie, M. & Fussenegger, M. Designing cell function: assembly of synthetic gene circuits for cell biology applications. Nat Rev Mol Cell Biol 19, 507-525, doi:10.1038/s41580- 018-0024-z (2018). 225 Doench, J. G. et al. Rational design of highly active sgRNAs for CRISPR-Cas9- mediated gene inactivation. Nat Biotechnol 32, 1262-1267, doi:10.1038/nbt.3026 (2014). 226 Hsu, P. D. et al. DNA targeting specificity of RNA-guided Cas9 nucleases. Nat Biotechnol 31, 827-832, doi:10.1038/nbt.2647 (2013). 227 Schlatter, S., Rimann, M., Kelm, J. & Fussenegger, M. SAMY, a novel mammalian reporter gene derived from Bacillus stearothermophilus alpha-amylase. Gene 282, 19- 31 (2002).

Page 193 of 200

228 Nguyen, N. T., He, L., Martinez-Moczygemba, M., Huang, Y. & Zhou, Y. Rewiring Calcium Signaling for Precise Transcriptional Reprogramming. ACS Synth Biol 7, 814- 821, doi:10.1021/acssynbio.7b00467 (2018). 229 Nguyen, A. W. & Daugherty, P. S. Evolutionary optimization of fluorescent proteins for intracellular FRET. Nat Biotechnol 23, 355-360, doi:10.1038/nbt1066 (2005). 230 Lee, S. M., Vasishtha, M. & Prywes, R. Activation and repression of cellular immediate early genes by serum response factor cofactors. The Journal of biological chemistry 285, 22036-22049, doi:10.1074/jbc.M110.108878 (2010). 231 Qi, L. S. et al. Repurposing CRISPR as an RNA-guided platform for sequence-specific control of gene expression. Cell 152, 1173-1183, doi:10.1016/j.cell.2013.02.022 (2013). 232 Aramburu, J. et al. Affinity-driven peptide selection of an NFAT inhibitor more selective than cyclosporin A. Science 285, 2129-2133 (1999). 233 Raikwar, N. S., Liu, K. Z. & Thomas, C. P. N-terminal cleavage and release of the ectodomain of Flt1 is mediated via ADAM10 and ADAM 17 and regulated by VEGFR2 and the Flt1 intracellular domain. PloS one 9, e112794, doi:10.1371/journal.pone.0112794 (2014). 234 Scheller, L., Strittmatter, T., Fuchs, D., Bojar, D. & Fussenegger, M. Generalized extracellular molecule sensor platform for programming cellular behavior. Nature chemical biology 14, 723-729, doi:10.1038/s41589-018-0046-z (2018). 235 Muller, M. et al. Designed cell consortia as fragrance-programmable analog-to-digital converters. Nat Chem Biol 13, 309-316, doi:10.1038/nchembio.2281 (2017). 236 Müller, M. et al. Designed cell consortia as fragrance-programmable analog-to-digital converters. Nature Chemical Biology 13, 309, doi:10.1038/nchembio.2281 (2017). 237 Zhou, S. et al. Targeted deletion of Smad4 shows it is required for transforming growth factor beta and activin signaling in colorectal cancer cells. Proc Natl Acad Sci U S A 95, 2412-2416 (1998). 238 Gimenez, C. A. et al. CRISPR-on system for the activation of the endogenous human INS gene. Gene Ther 23, 543-547, doi:10.1038/gt.2016.28 (2016). 239 Senior, M. Novartis signs up for Google smart lens. Nat Biotechnol 32, 856, doi:10.1038/nbt0914-856 (2014). 240 Selvaraj, N. in Healthcare Innovation Conference (HIC), 2014 IEEE. 83-86.

Page 194 of 200

241 Selvaraj, N. & Narasimhan, R. Automated prediction of the apnea-hypopnea index using a wireless patch sensor. Conference proceedings : ... Annual International Conference of the IEEE Engineering in Medicine and Biology Society. IEEE Engineering in Medicine and Biology Society. Annual Conference 2014, 1897-1900, doi:10.1109/EMBC.2014.6943981 (2014). 242 Chan, A. M., Ferdosi, N. & Narasimhan, R. Ambulatory respiratory rate detection using ECG and a triaxial accelerometer. Conference proceedings : ... Annual International Conference of the IEEE Engineering in Medicine and Biology Society. IEEE Engineering in Medicine and Biology Society. Annual Conference 2013, 4058-4061, doi:10.1109/EMBC.2013.6610436 (2013). 243 Stauffer, F. et al. Skin Conformal Polymer Electrodes for Clinical ECG and EEG Recordings. Adv Healthc Mater 7, e1700994, doi:10.1002/adhm.201700994 (2018). 244 Kulkarni, G. S., Reddy, K., Zhong, Z. & Fan, X. Graphene nanoelectronic heterodyne sensor for rapid and sensitive vapour detection. Nat Commun 5, 4376, doi:10.1038/ncomms5376 (2014). 245 Garbarino, M., Lai, M., Bender, D., Picard, R. W. & Tognetti, S. in Wireless Mobile Communication and Healthcare (Mobihealth), 2014 EAI 4th International Conference on. 39-42. 246 Tybrandt, K. et al. High-Density Stretchable Electrode Grids for Chronic Neural Recording. Adv Mater 30, e1706520, doi:10.1002/adma.201706520 (2018). 247 Renz, A. F., Reichmuth, A. M., Stauffer, F., Thompson-Steckel, G. & Voros, J. A guide towards long-term functional electrodes interfacing neuronal tissue. J Neural Eng 15, 061001, doi:10.1088/1741-2552/aae0c2 (2018). 248 Bai, P. et al. A fully human transgene switch to regulate therapeutic protein production by cooling sensation. Nature medicine 25, 1266-1273, doi:10.1038/s41591-019-0501-8 (2019). 249 Nussinovitch, U. & Gepstein, L. Optogenetics for in vivo cardiac pacing and resynchronization therapies. Nat Biotechnol 33, 750-754, doi:10.1038/nbt.3268 (2015). 250 Bruegmann, T. et al. Optogenetic control of heart muscle in vitro and in vivo. Nat Methods 7, 897-900, doi:10.1038/nmeth.1512 (2010). 251 Carmo, M., Fritz, D. L., Merge, J. & Stolten, D. A comprehensive review on PEM water electrolysis. Int J Hydrogen Energ 38, 4901-4934, doi:10.1016/j.ijhydene.2013.01.151 (2013).

Page 195 of 200

252 in Neuroscience (ed George J Augustine Dale Purves, David Fitzpatrick, Lawrence C Katz, Anthony-Samuel LaMantia, James O McNamara, and S Mark Williams) Ch. 3, (Sinauer Associates, 2001). 253 Rios, E. Calcium-induced release of calcium in muscle: 50 years of work and the emerging consensus. J Gen Physiol 150, 521-537, doi:10.1085/jgp.201711959 (2018). 254 Ye, H. & Steiger, A. Neuron matters: electric activation of neuronal tissue is dependent on the interaction between the neuron and the electric field. J Neuroeng Rehabil 12, 65, doi:10.1186/s12984-015-0061-1 (2015). 255 Joucla, S., Gliere, A. & Yvert, B. Current approaches to model extracellular electrical neural microstimulation. Front Comput Neurosci 8, 13, doi:10.3389/fncom.2014.00013 (2014). 256 Bardou, A. L. et al. Directional variability of stimulation threshold measurements in isolated guinea pig cardiomyocytes: relationship with orthogonal sequential defibrillating pulses. Pacing and clinical electrophysiology : PACE 13, 1590-1595 (1990). 257 Takayama, Y., Wagatsuma, A., Hoshino, T. & Mabuchi, K. Simple micropatterning method for enhancing fusion efficiency and responsiveness to electrical stimulation of C2C12 myotubes. Biotechnol Prog 31, 220-225, doi:10.1002/btpr.2003 (2015). 258 Wagenaar, D. A. & Potter, S. M. A versatile all-channel stimulator for electrode arrays, with real-time control. J Neural Eng 1, 39-45, doi:10.1088/1741-2560/1/1/006 (2004). 259 Muller, J. et al. High-resolution CMOS MEA platform to study neurons at subcellular, cellular, and network levels. Lab Chip 15, 2767-2780, doi:10.1039/c5lc00133a (2015). 260 Marsland, B. J. & Kopf, M. T-cell fate and function: PKC-theta and beyond. Trends Immunol 29, 179-185, doi:10.1016/j.it.2008.01.005 (2008). 261 Morgan, R. A. et al. Case report of a serious adverse event following the administration of T cells transduced with a chimeric antigen receptor recognizing ERBB2. Mol Ther 18, 843-851, doi:10.1038/mt.2010.24 (2010). 262 Chmielewski, M. & Abken, H. TRUCKs: the fourth generation of CARs. Expert Opin Biol Ther 15, 1145-1154, doi:10.1517/14712598.2015.1046430 (2015). 263 Roybal, K. T. et al. Precision Tumor Recognition by T Cells With Combinatorial Antigen-Sensing Circuits. Cell 164, 770-779, doi:10.1016/j.cell.2016.01.011 (2016). 264 Schmidts, A. & Maus, M. V. Making CAR T Cells a Solid Option for Solid Tumors. Front Immunol 9, 2593, doi:10.3389/fimmu.2018.02593 (2018).

Page 196 of 200

265 Daringer, N. M., Dudek, R. M., Schwarz, K. A. & Leonard, J. N. Modular extracellular sensor architecture for engineering mammalian cell-based devices. ACS Synth Biol 3, 892-902, doi:10.1021/sb400128g (2014). 266 Cebrian-Serrano, A. & Davies, B. CRISPR-Cas orthologues and variants: optimizing the repertoire, specificity and delivery of genome engineering tools. Mamm Genome 28, 247-261, doi:10.1007/s00335-017-9697-4 (2017). 267 Gemmill, D., D'Souza, S., Meier-Stephenson, V. & Patel, T. R. Current approaches for RNA-labelling to identify RNA-binding proteins. Biochem Cell Biol 98, 31-41, doi:10.1139/bcb-2019-0041 (2020). 268 Sajwan, S. & Mannervik, M. Gene activation by dCas9-CBP and the SAM system differ in target preference. Sci Rep 9, 18104, doi:10.1038/s41598-019-54179-x (2019). 269 Josipovic, G. et al. Antagonistic and synergistic epigenetic modulation using orthologous CRISPR/dCas9-based modular system. Nucleic Acids Res 47, 9637-9657, doi:10.1093/nar/gkz709 (2019).

Page 197 of 200

Page 198 of 200

Acknowledgements

I would like to thank: - Prof. Dr. Martin Fussenegger the opportunity to conduct my research in his laboratory and for the support that he gave me during that time, - Prof. Dr. Petra S. Dittrich, Prof. Dr. János Vörös, and Prof. Dr. Randall Platt for agreeing to be my co-examiners, - Prof. Dr. Niko Beerenwinkel for agreeing to be the chair or the examination committee, - Peter Buchmann for his great support with electronic engineering, - Leo Scheller for being great to work with, helpful discussions, pleasant coffee breaks, and for help with translating the summary into German, - Pratik Saxena for his generosity in sharing his expertise, advice, suggestions, and for comments on this thesis, - David Fuchs, as he deserves to be acknowledged in every thesis defended in this department, - Marc Folcher for his advice, especially on implant construction, - Marie-Didiée Hussherr for her help with animal experiments in Basel, - Mingqi Xie, Shuai Xue and Jiawei Shao for their help with animal experiment in Shanghai, - Prof. Dr. Haifeng Ye and his group for hosting me in Shanghai, - Ghislaine Charpin-El-Hamri and Marie Daoud El-Baba for their help with animal experiments in Lyon, - Sebastian Bürgel for his help at the beginning of the electrogenetic project, - The members of lab 5.44 (Pratik, Ying, Peng, Daniel, Mingqi, Taeuk, Ryosuke, and Anton) for their collegiality, - Collaborators from Single Cell Unit: Telma Lopes, Verena Jäggin and Ola Gumienny for their support using FACS, - Colleagues from D-BSSE who supported me with advice, help, or materials, - All members of Prof. Fussenegger’s laboratory during my time there (Viktor, Toby, Pascal, Rym, Hui, Ferdi, Marius, Maysam, Ana, Bozho, et al.), - My family for their unconditional lifetime support, - Finally, my wonderful fiancée Lisa who has tirelessly supported me.

Page 199 of 200

Page 200 of 200