Ontogeny of the Intestinal Circadian Clock and Its Role in the Response to Clostridium Difficile Toxin B

Ontogeny of the intestinal circadian clock and its role

in the response to Clostridium difficile toxin B

A dissertation submitted to the

Graduate School

of the University of Cincinnati

in partial fulfilment of the

requirements for the degree of

Doctor of Philosophy

In the Department of Pharmacology &

Systems Physiology of the College of Medicine

Andrew Rosselot

B.S. Biology, Wittenberg University

October 2019

Committee Chair: Christian I. Hong Ph.D. Abstract:

The endogenous clock of the intestine regulates physiological processes ranging from nutrient absorption to the pathogenic response. The developmental timepoint when the human intestinal clock becomes active is unknown. We investigated intestinal circadian clock ontogeny using in vitro samples that are representative of distinct developmental timepoints. Induced pluripotent stem cells (iPSCs) were differentiated into 3D human intestinal organoids (HIOs) to mimic intestinal embryonic development in vitro. HIOs were then matured beyond their early fetal state via kidney capsule transplantation. Differentiation of iPSCs into HIOs did not activate robust circadian clock activity. Enteroids isolated from kidney capsule matured HIOs possessed a functional circadian clock, similar to adult biopsy derived human intestinal enteroids (bHIEs).

Samples were challenged with toxin B (TcdB) from Clostridium difficile to provide functional insights on intestinal clock activity. The necrotic cell death response to TcdB was clock phase- dependent in samples that possessed an active clock and anti-phasic between mouse enteroids and bHIEs. RNA-seq analysis of mouse enteroids and bHIEs showed both possess robust rhythmic gene expression with up to 20% and 8% of their transcriptome oscillating, respectively.

The phase and identity of rhythmic genes was however species-dependent. Interestingly, we found Rac1 to be the only TcdB target rhythmically expressed in both mouse enteroids and bHIEs.

Further, Rac1 expression was anti-phasic between mouse and human samples. In this thesis we have characterized intestinal circadian clock ontogeny using novel 3D in vitro intestinal models.

We utilized the characterization to show for the first time that a functional clock is required for a circadian phase-dependent response to Clostridium difficile toxin B. Rhythmic Rac1 expression was anti-phasic between mouse and human enteroids which correlated with their anti-phasic necrotic cell death response to TcdB. These findings underscore the use of the human intestinal organoids/enteroids for understanding fundamental human intestinal circadian biology and its translatable role in promoting circadian phase-dependent human fitness.

iii

Acknowledgements:

Thank you to my wife, Hilary. Your support for me throughout my time in graduate school has been ceaseless. During the many peaks and troughs of grad school you were always there to keep me pushing towards this moment. I may have been able to accomplish this goal without your support but luckily didn’t have to nor would I have wanted to. I’m excited to start my “big kid” life with you.

Thank you to my family, immediate, extended and non-blood/law. Doing my doctorate at UC allowed me to be close to you all while working towards this accomplishment. Although I don’t have a control to compare against, I have highly valued your proximity while working towards this goal. The respite and memories made during this past 5-years are something I will always be grateful for.

Thank you, Chris Hong. As my advisor you always challenged me to think critically and independently. Your high expectations shaped an excellent, valuable training environment, and I am a better scientist for it.

Thank you, Toru Matsu-ura. This work would not have been possible without you. I will forever be grateful for your patience and diligence while training me in the many cell biology tasks I didn’t know at the beginning of my grad career.

Thank you to all my committee members, Sean Moore, Alison Weiss, Jim Wells and Yana

Zavros. From my qualifying exam to final committee meeting, each member showed a great amount of interest and support for my work. Your feedback was crucial in helping me create the project detailed in the following chapters.

Thank you to the following lab’s and lab members for your assistance, both scientific and technical.

The lab of Chris Hong – University of Cincinnati:

Miri Park

Suengwon Lee

Krithika Ramasamy-Subramanian

Kaoru Matsu-ura

Mokryun Baek

The lab of Jim Wells – Cincinnati Children’s Hospital Medical Center:

Taylor Broda

Heather McCauley

Lauren Haines

Jorge Munera

The lab of Michael Helmrath – Cincinnati Children’s Hospital Medical Center:

Nambirajan Sundaram

Jennifer Hawkins

The lab of Alison Weiss – University of Cincinnati

Suman Pradhan

The lab of John Hogenesch – Cincinnati Children’s Hospital Medical Center:

Lauren Francey

Gang Wu

Danilo Flores

Robert Schmidt

The lab of Yana Zavros – University of Cincinnati

Jayati Chakrabati

The lab of Garret FitzGerald – University of Pennsylvania

Guangrui Yang

The lab of Noah Shroyer – Baylor University

University of Cincinnati Live Microscopy Core

Chet Closson

Table of Contents:

Abstract………………………………………………………………………………………………….…ii

Acknowledgements………………………………………………………………………………………iv

Figures/Tables...………………………………………………………………………………………..…x

Abbreviations……………………………………………….…………………………………………….xi

Chapter 1: Introduction…………………………………………………………………………………..1

1.1 Conservation of the molecular clock across eukaryotes……………………………………1

1.2 Circadian misalignment due to modern lifestyle and its implications in disease

pathogenesis………………………………………………………………………………….…6

1.3 The mammalian central and peripheral circadian clocks: same clock different story...…7

1.4 Human intestinal organoids and enteroids as models for extending our understanding of

the endogenous clock of the intestine...…………………………………………...………..12

1.5 General introduction to Clostridium difficile pathogenesis………………………...……...17

1.5.1 Spore formation and transmission…………………………………………………...17

1.5.2 The Clostridium difficile pathogenicity locus………………………………...…...…18

1.5.3 Toxin B cell receptors…………………………………………………………………18

1.5.4 Mechanism of action for TcdA-/TcdB-mediated cell death………………………..20

Chapter 2: Ontogeny of the intestinal circadian clock……………………………………………….25

2.1 Abstract……....……………………………………………………………………………..….25

2.2 Introduction……………...……………………………………………………………….....….25

2.3 Methods…………………………………………………………………………………..…….31

2.3.1 HIO generation………………………………………………………………………...31

2.3.2 Generation and maintenance of kidney capsule-matured- and biopsy derived-

human intestinal enteroids……………………………………...... 32

2.3.3 Bmal1-luciferase lentiviral transduction and monitoring of Bmal1 activity……….33

vii

2.3.4 Timecourse sample collection………………………………………………………..34

2.3.5 Statistical analysis……………………………………………………………………..37

2.3.6 Tables for reagents and primers……………………………………..………………37

2.4 Results………………………………………………………………………………………….40

2.4.1 Bmal1-luc rhythms are not robust in HIOs………………….……………………....40

2.4.2 In vivo maturation of HIOs prompts GI circadian clock development…………....43

2.5 Conclusions and discussion………………………………………………………………….44

Chapter 3: The intestinal circadian clock regulates a phase-dependent response to toxin B from

Clostridium difficile………………………………………………………………………….………...…46

3.1 Abstract…………………………………………………………………………………………46

3.2 Introduction………………………………………………………………………………….…47

3.3 Methods………………………………………………………….……………………………..50

3.3.1 Animals……………….………………………………………………………………...50

3.3.2 Isolation and maintenance of mouse enteroids…….………………………………51

3.3.3 Experiment design to test for a circadian phase-dependent response to TcdB...52

3.3.4 RNA-sequencing analysis…………………………………………………………….54

3.4 Results……………………………………………………………………………………….…56

3.4.1 Intestinal 3D cultures with a functional clock have a circadian phase-dependent

response to C. diff Toxin B……..……………………………………………………..56

3.4.2 Mouse and human enteroids possess transcriptome wide rhythmic gene

expression and out-of-phase expression of core clock genes and Rac1…………62

3.5 Conclusions and discussion………………………………………………………………….65

Chapter 4: General discussion and future directions…………………………………………………68

4.1 Identification of circulating circadian signals that establish circadian clock function in the

developing intestine………………………………………………………………………...…69

4.1.1 In vitro activation of the HIO circadian clock….……………………………………...69

viii

4.1.2 Evaluation of host circadian input on clock activation in transplanted HIOs …....71

4.2 Characterization of intestinal clock aging and its influence on the response to Clostridium

difficile toxin B…………..………………………………………………………………….…..75

4.3 Summary……………………………………………………………………………………….78

References……………………………………………………………………………………………….82

Appendix…………………………………………………………………………………………………94

Figures/Tables:

Figure 2.1 Robust circadian clock activity is not present throughout PSC differentiation into HIOs. Figure 2.2 HIOs have low, non-circadian, expression of core clock genes. Figure 2.3 Overview of the source and method to establish in vitro 3D intestinal organoids and enteroids. Figure 2.4 HIO transplantation activates the circadian clock. Figure 2.5 A fully functional TTFL is not present in kcHIEs. Table 2.1 Key resources table. Table 2.2 List of primers.

Figure 3.1 Bi-directional interactions between the circadian clock, microbiota and immune system. Figure 3.2 TcdB phase-response experiment design. Figure 3.3 Synchronization of PER2::LUCIFERASE mouse enteroids 12-hours apart drives samples out-of-phase. Figure 3.4 Necrotic, not apoptotic, cell death is phase-dependent in PER2::LUC control mouse enteroids. Figure 3.5 Tamoxifen treated Bmal1Fx/Fx-EsrCRE mouse enteroids lack rhythmic core clock gene expression. Figure 3.6 Tamoxifen treated arrhythmic Bmal1Fx/Fx-EsrCRE mouse enteroids do not respond to TcdB in a circadian phase-dependent manner. Figure 3.7 Necrotic cell death is phase-dependent in human 3D intestinal models with a functional clock. Figure 3.8 Mouse enteroids possess transcriptomic rhythms. Figure 3.9 Human biopsy derived enteroids possess transcriptomic rhythms. Figure 3.10 Rhythmic transcript phase, in enteroids, is species-dependent.

Figure 4.1 Environmental manipulations to evaluate host circadian input on clock development in transplanted HIOs. Figure 4.2 Network interaction between the core circadian clock loop and SIRT1/NAD+. Figure 4.3 Preliminary insights on the influence of donor age on the bHIE circadian phase-dependent response to TcdB.

Abbreviations:

3D 3-dimensional AL ad libitum AVP arginine vasopressin bHIE biopsy human intestinal enteroid bHLH basic helix loop helix BMAL1 brain and muscle arnt like protein-1 BMDM bone-marrow derived macrophages CCG circadian clock gene CCHMC cincinnati children’s hospital medical center CDI clostridium difficile infection CLOCK circadian locomotor output cycles kaput CPD cysteine protease domain CRD cysteine rich domain CSPG4 chondroitin sulfate proteoglycan-4 D36/D24 dexamethasone-36 / dexamethasone-24 DD constant darkness DE definitive endoderm DEX dexamethasone DLMO dim-light melatonin onset EGF epidermal growth factor ESC embryonic stem cell FAA food anticipatory activity FBS fetal bovine serum FEO food entrainable oscillator FZD frizzled GABA gamma-amminobutyric acid GTD glucosyl transferase domain HG hindgut HIO human intestinal organoid IEC intestinal epithelial cell IGF-1 insulin-growth factor-1 ipRGC intrinsically photosensitive retinal ganglion cells iPSC induced pluripotent stem cell ISC intestinal stem cell ISEMF intestinal sub-epithelial myofibroblasts kcHIE kidney-capsule human intestinal enteroid KD knockdown KO knockout LD light-dark LL constant light LPS lipopolysaccharide mTORC1 mammalian target of rapamycin complex 1 NAC n-acetyl cysteine NAM nicotinamide NAMPT nicotinamide phosphoribosyltransferase NAT n-acetyltransferase NMN nicotinamide mononucleotide NOB nobiletin

NSG non-obese diabetic severe combined immunodeficiency interleukin-2rγnull PaLoc pathogenicity locus PAS per-arnt-sim PC paneth cell PSC pluripotent stem cell PTO post-translational oscillator PVRL3 poliovirus receptor-like 3 qRT-PCR quantitative real-time polymerase chain reaction RF restricted feeding RHT retinohypothalamic tract RORE retinoic acid nuclear orphan receptor element SCN suprachiasmatic nucleus TA transit amplifying TcdB clostridium difficile toxin B TLR toll-like receptor TOD time-of-day TPM transcripts per million TTFL Transcriptional-translational feedback loop TTO transcriptional-translational oscillator VIP vasoactive intestinal peptide WCC white collar complex

xii

1. Introduction

1.1 Conservation of the molecular clock across eukaryotes

Throughout an organism’s lifespan, its biology is constantly changing via developmental and aging mechanisms while also responding to environmental stimuli. Often, the adaptive change or response to an external stimulus occurs at discrete timepoints. Environmental light- dark (LD) cycles are however a chronic environmental change that repeat approximately every

24-hours. Nearly every species on Earth experience these cyclical LD cycles and until recently, through the advent of artificial light, much of human activity coincided with the presence of natural light. It is therefore not surprising that mechanisms are present to coordinate biological processes to LD cycles.

Drosophila eclosion was the first biological process observed to occur in a time dependent fashion, taking place in the morning to coincide with ideal conditions for pupae survival1.

Seventeen years later the first molecular insights into what drives eclosion rhythms were provided.

Drosophila mutant flies with arrhythmic, short- and long-period eclosion and locomotor activity were generated. Each mutation mapped to the same genetic locus, implicating a single gene, termed Period (Per), in driving the phenotypes2. Analysis of the Period gene identified the diurnal rhythm of its transcriptional product and the loss of the rhythm in Period mutant (Per0) flies3. When the DNA sequence encoding the Per RNA fragment was re-introduced into Per0 flies, rhythmic eclosion and locomotor activity was reinitiated4. Thus, the first circadian clock gene, Period, was mapped and shown to functionally produce rhythmic RNA transcription/translation that drives

Drosophila circadian phenotypes. These foundational insights propelled a new line of research into molecular biological rhythms across multiple organisms that include: cyanobacteria,

Neurospora crassa, Drosophila and mammals.

Cyanobacteria are light sensing prokaryotes and the simplest organism known to present circadian clock driven oscillations. A bi-product of cyanobacteria photosynthesis is oxygen.

Nitrogenase, a key enzyme in nitrogen fixation, cannot function in the presence of oxygen. The ability of cyanobacteria to carry out photosynthesis and nitrogen fixation was therefore perplexing to biologists because of their mutual exclusivity. Temporal separation of these two processes was proposed to explain the phenotype. Investigation into this hypothesis found that cyanobacteria nitrogenase activity peaks at night5 and photosynthesis occurs during the day6.

Inherent circadian clock driven, rather than light-dark cycle driven, nighttime nitrogen fixation was shown with maintenance of nitrogenase activity during the subjective night under constant light exposure5,6. Ishiura and colleagues were the first to fully characterize the influence of the kaiABC gene cluster on cyanobacteria clock function7. Mutation of kaiA or kaiC, but not kaiB, induced arrhythmicity. kaiA overexpression augmented kaiBC promoter activity. Temporal resetting of kaiC expression promoted individual cyanobacteria to phase-reset and kaiC overexpression resulted in global rhythm repression7. Transcriptionally driven cyanobacteria clock function aligned with concurrent findings of a transcriptional-translational feedback loop (TTFL) driving molecular circadian rhythms of Drosophila, Neurospora crassa and mammals8. However, follow- up studies found cyanobacteria with constitutive kaiBC9 or lack of kaiABC transcriptional activity10 still possessed circadian rhythms. Rhythmic post-translational modifications were subsequently shown to be a primary contributor to the cyanobacteria rhythms. Cyanobacteria post-translational rhythms are centered on KaiC phosphorylation at Ser431 and Thr43211 with KaiA and KaiB regulating KaiC autophosphorylation activity. KaiA promotes KaiC autophosphorylation12 and is sequestered from further activation by KaiB13. Phosphorylation at Ser431 or Thr432 dictates cyanobacteria phase. When phosphorylation of KaiC at Thr432 is high, KaiA will promote KaiC autokinase activity that results in higher quantities of stable KaiC with double phosphorylated

Ser431/Thr432. Dephosphorylation of the Ser431-Thr432 KaiA-KaiC complex into Ser431 mono- phosphorylated KaiA-KaiC prompts binding of KaiB that will push KaiC into an inactive unphosphorylated form14. These data generated a post-translational oscillator (PTO) centric view of the cyanobacteria clock. Teng et al. (2013) however showed the PTO and transcriptional-

translational oscillator (TTO) were indispensable of one another for rhythmic cyanobacteria function15. When TTO activity is constitutive, cyanobacterial population scale rhythms become desynchronized, under constant conditions. When the PTO was held constant, molecular rhythms were lost. These data support the necessity of both oscillators: PTO for the generation and sustainment of rhythms and the TTO for the maintenance of population rhythms under constant conditions15. The above work has provided a deep understanding of the cyanobacterial clock, the earliest known evolutionary oscillator. Interestingly, the molecular interaction between the

TTO/PTO driving cyanobacteria oscillations is quite different from the TTFL circadian oscillator found in higher organisms. Somewhere along the evolutionary timeline the TTO/PTO paradigm switched to a TTFL oscillator and Neurospora crassa has been used as the simplest/earliest evolutionary model to uncover the molecular mechanisms of the TTFL loop.

Molecular circadian clock activity of Neurospora, Drosophila and mammals all function via a conserved TTFL8,16. Neurospora crassa is a filamentous fungus that exhibits conidiation oscillations with a period length of ≈22-hours. The first Neurospora circadian mutants were described by Feldman and Hoyle (1973) and termed frequency mutants17. Subsequent work successfully isolated and cloned frequency (frq) and showed that the mutated clock phenotype was alleviated when frq was reintroduced into circadian mutant strains18. The activation arm of the Neurospora TTFL consists of two light responsive factors, White collar-1 (Wc-1) and White collar-2 (Wc-2)19. WC-1 and WC-2 interact with one another via PER-ARNT-SIM (PAS) domains20 to form the white-collar complex (WCC). The WCC activates frq via Zn finger domain interaction with the frq clock box (C-box) promotor region21. The binding of WC-1 on the frq c-box is not rhythmic. Clockswitch-1 (CSW-1) rhythmically induces a closed chromatin structure at the c-box promoter of frq, providing a limited and rhythmic temporal window when WC-2 can bind and form the WCC22. FRQ binds with FRQ-interacting RNA helicase and translocates into the nucleus to inhibit the WCC23. FRQ inhibition of the WCC is in part mediated by removal of WC-1 from the nucleus into the cytoplasm to be degraded24. Attenuated frq transcription coupled with continued

phosphorylation and degradation of FRQ releases the WCC from FRQ inhibition, freeing the positive arm to become active at the frq promoter to restart the loop. Neurospora crassa has provided an excellent, cost effective model system for studying the molecular mechanisms that drive circadian clock activity. Initial progress in understanding the clock using Neurospora crassa was paralleled by simultaneous efforts to understand the clock within Drosophila. The convergence and validation of information coming from both models was instrumental in driving a deep understanding of the eukaryotic molecular clock.

The Drosophila TTFL consists of the core clock elements: dClock (dClk), Cycle (Cyc), dPeriod (dPer) and Timeless (Tim). dCLK and CYC are analogous to WC-1 and WC-2 of

Neurospora. Both are PAS containing transcriptional activators that heterodimerize to form the positive arm of the Drosophila clock25,26. Drosophila clock-regulated targets are activated by bHLH mediated dCLK:CYC activity at their E-box regulatory elements27. dCLK:CYC binding to the E- box of dPer and Tim drive their rhythmic transcription28. Within the cytoplasm dPER is stabilized by DOUBLETIME (DBT) via phosphorylation29. Phosphorylated dPER then forms a heterodimer with TIM to promote further phosphorylation and nuclear translocation of dPER30,31. Nuclear dPER complexes with DBT32 to traffic DBT to dCLK resulting in phosphorylation of dCLK and repression of dCLK:CYC transcriptional activity33,34. Similar to FRQ in Neurospora, dPER negative feedback on dCLK:CYC results in reduced transcription of repressive elements, dPer and tim, which, coupled with the degradation of dPER, releases dCLK from its repressive state to restart the circadian cycle. Unlike Neurospora, Drosophila possesses a secondary retinoic-acid nuclear orphan receptor element (RORE) circadian feedback loop. The RORE loop includes Vrille (Vri) and Par domain protein 1 ϵ (Pdp1ϵ). dCLK:CYC regulate the rhythmic expression of Vri and

Pdp1ϵ35,36 that feedback negatively or positively on dClk, respectively37. The Drosophila circadian system is separated into two categories, central and peripheral. The brain of Drosophila contains the central clocks that are subdivided into multiple neuronal projections38. Outside the brain are the Drosophila peripheral clocks, present within the wings39, intestine40 and Malphigian tubules41,

among other tissues42. How Drosophila central and peripheral clocks are connected/signal to one another is still not completely understood. Classically, the mammalian circadian system is seen as a top-down hierarchy with central clocks organizing clock function in non-photosensitive peripheral clocks. Drosophila peripheral organs are however directly light responsive39 which could mitigate the need for central clocks to synchronize peripheral clock function to an LD cycle.

The potential variability in peripheral clock coordination to an LD cycle underscore subtle differences that likely exist between circadian function in Drosophila and mammal. Direct investigation of the mammalian circadian system is required to offer the most comprehensive and translatable insights into human circadian clock function and how it in turn impacts human health.

The mammalian circadian clock is present ubiquitously throughout the body, providing temporal information to cellular and physiological processes ranging from metabolism to cell proliferation. Molecularly, circadian rhythms are generated via the analogous TTFL mechanism found in Neurospora and Drosophila. Two basic helix-loop-helix (bHLH) PER-ARNT-SIM (PAS) domain containing proteins, CLOCK (circadian locomotor output cycles kaput) and BMAL1 (brain and muscle Arnt-like protein-1), form a heterodimeric circadian transcription factor that recognizes

E-box motifs on target gene promoters including Period (Per1, Per2, Per3) and Cryptochrome

(Cry1, Cry2). PER and CRY bind together in the cytoplasm43 and translocate into the nucleus where they interact with the CLOCK-BMAL1 heterodimer to repress its activity44. PER and CRY are then ubiquitinated and degraded, relieving CLOCK-BMAL1 from negative feedback to re- initiate the cycle45,46. Orphan nuclear receptor transcription factors also impact molecular clock activity via regulation of Bmal1, analogous to the secondary RORE loop described in Drosophila.

BMAL1:CLOCK drive rhythmic expression of positive (Rorα, Rorβ, Rorγ) and negative (Rev-erbα and Rev-erbβ) nuclear orphan receptors that competitively bind to ROREs on Bmal147–49. Activity of core clock elements outside of the TTFL loop prompt ≈50% of protein coding genes to oscillate somewhere in the body50 which results in tissue-specific circadian physiological functions. Like

Drosophila, the mammalian circadian clock is subdivided into central and peripheral clocks that

are described in greater detail in section 1.3. Evolutionarily, the mechanisms above have operated unperturbed for generations and likely continue to do so in cyanobacteria, Neurospora, Drosophila and non-human mammals. Recent technological advancements have, however, provided humans the unique capacity to easily disrupt their circadian system. We are just beginning to understand the negative effects of acute and chronic circadian disruption on human health.

1.2 Circadian misalignment due to modern lifestyle and its implications in

disease pathogenesis

Throughout human history, environmental changes capable of rapidly desynchronizing the circadian clock were not present. The advent of artificial light in the early 20th century allowed the night to become easily and continuously illuminated. Concurrently, airline travel offered humans the opportunity to rapidly cross multiple time zones, producing jet-lag. Desynchronization of our circadian clocks short term is uncomfortable, causing daytime fatigue and nighttime alertness, but is not lethal. Continuous desynchronization negatively impacts human health. Cardiovascular and metabolic disease have been epidemiologically correlated with rotating shift work51,52. A confounding variable in epidemiological analyses of shift workers and circadian desynchronization is however the number of hours slept. Shift workers sleep fewer hours prior to working compared to non-shift workers. A lack of sleep could be the causal between negative health effects and shift work. A recent study did find the association between shift work and metabolic syndrome is attenuated when accounting for sleep duration53. Lab-controlled studies with human subjects, however, support the direct role of circadian desynchronization in causing metabolic disease. Desynchronization of circadian rhythms results in attenuated leptin levels, inverted cortisol rhythms, heightened mean arterial blood pressure and post-prandial hyperglycemia. Circadian desynchronization does not influence post-prandial insulin levels54, indicating desynchronization induced hyperglycemia is the result of insulin insensitivity. The

confounding role of sleep was directly addressed in a separate lab-controlled study where subjects were either sleep deprived or sleep deprived and circadian desynchronized. Coupling sleep deprivation and circadian desynchronization resulted in lower insulin sensitivity compared to sleep deprivation alone55. These data indicate jobs that require chronic circadian desynchronization may predispose workers to poor health outcomes as a result of continual disruption of their circadian system.

The circadian phase of the suprachiasmatic nucleus (SCN), the master mammalian clock, is set by LD cycles56. The SCN relays LD cycle information to peripheral tissues to orchestrate their oscillations with the LD cycle. SCN synchronization of peripheral clocks can be interrupted by food intake, another phase-setting signal for peripheral oscillators57. When food consumption is restricted to atypical times, the light phase in nocturnal rodents, SCN and food intake signals lose coordination, resulting in altered rhythmicity of metabolic genes in the liver58. In the above studies, timing of food availability was not controlled, making it impossible to know whether the phenotypes found were a product of altered light schedule on the SCN, altered food intake on peripheral clocks or a combination of both. Although central and peripheral oscillators operate under the same TTFL loop, their entrainment cues, outputs and role within the whole-body circadian system differ. A deeper understanding of central and peripheral oscillator function under normal and desynchronized conditions is required to understand the role either has in promoting negative health outcomes during chronic circadian disruption.

1.3 The mammalian central and peripheral circadian clocks: same clock different

story

The Suprachiasmatic nucleus (SCN) is located in the hypothalamus and serves as the master clock of the circadian system. LD signals are perceived by melanopsin containing intrinsically photosensitive retinal ganglion cells (ipRGCs)56,59,60 of the eye, and sent to the SCN

via the retinohypothalamic tract (RHT)61 to coordinate SCN oscillations to the LD cycle. As a population, SCN neurons produce rhythmic waves of activity that progress from its dorsal to ventral region62. Individually, each of the SCN’s 20,000 neurons contain a cell-autonomous clock.

The firing rate of individual neurons, an indirect measure of its autonomous clock, in vitro is heterogenous; neurons within a single culture can have firing activity with short, circadian and long period lengths or be arrhythmic63,64. When co-cultured at low density, SCN neurons are incapable of synchronizing their firing activity to one another65, but will synchronize when plated at high density66. The lack of synchrony in low-density vs. high-density neuronal cultures shows neurons can communicate intercellularly but only over short distances. Both gamma-aminobutyric acid (GABA)63 and vasoactive intestinal peptide (VIP)66 have been implicated as intercellular signaling molecules that synchronize SCN neurons67. Regardless of the intercellular signaling factor, the mechanism prompting synchronized circadian waves across the SCN has been difficult to ascertain from analysis of neuronal activity alone. It is becoming increasingly apparent that

SCN circadian activity is not only a product of the autonomous neuron clocks but also SCN astrocyte clock activity.

Three recent studies have established that SCN astrocytes are a key cell type influencing the synchronization and function of the SCN neurons and thus whole-body rhythms. Behavioral activity rhythms in astrocyte specific Bmal1KO mice are either bimodal68 or lengthened69,70. To show direct input of the astrocyte clock on the clock of neurons, synchronized control and synchronized siRNA Bmal1KD astrocytes were co-cultured with non-synchronized SCN neurons.

Control astrocyte co-culture produced neurons with circadian activity whereas Bmal1KO’s did not.

The co-culture neuronal synchronization was lost via inhibition of GABAA, implicating GABA as an intercellular signaling factor between astrocytes to neurons68. Astrocyte mediated glutamate signaling also regulates the astrocyte-neuron connection within the SCN70. Astrocytes and neurons oscillate out-of-phase to one another within the dorsal region of the SCN. Astrocytes oscillate in-phase with the rhythm in extracellular glutamate. Inhibition of glutamate synthase, the

enzyme responsible for glutamate to glutamine conversion within astrocytes, prompted heightened extracellular glutamate, due to excess levels of glutamate within the cytosol of astrocytes and also resulted in a loss of clock activity. Clock was also lost specifically in the dorsal

SCN when the glutamate receptor NR2C was inhibited70. These data extended upon the findings by Barca-Mayo et al. (2017) showing that glutamate release from astrocytes signals to SCN neurons which prompts GABA release into the extracellular space68. Specifically, the release of

GABA and inhibition of action potential firing from neurons occurs during the inactive phase and explains why astrocytes and neurons are anti-phasic70. These data from astrocytes indicate that

SCN signaling during the active phase is a result of light sensitive VIPergic neuronal function 71 that becomes inhibited by the release of glutamate from astrocytes that prompts increased extracellular GABA during the inactive phase. Rhythmic neuronal activity of the SCN orchestrates the peripheral oscillators found throughout the body72 via humoral73,74 and neuronal signaling75,76.

The classical view of the circadian system is hierarchical, the SCN receives LD cues and relays that information to peripheral clocks throughout the body. Under this model, the peripheral clocks are dependent upon the SCN, unable to synchronize their oscillations without SCN input. Nutrient consumption is, however, a robust entrainment cue for peripheral oscillators57 and liver specific

Bmal1 knockout prompted the loss of rhythmic gene expression within the liver but not in other tissues77. Tissue specific clock function is therefore not entirely SCN dependent and inter-tissue variability in rhythmic gene expression is beginning to be understood50

Endogenous circadian clocks are present in each tissue of the body, driving tissue specific physiological rhythms that are entrained by cues originating from either the SCN or nutrient intake.

Under normal conditions the SCN and nutrient intake work in concert to promote robust peripheral clock rhythms that are in-phase with the SCN. The majority of activity and food consumption in mice occurs during the dark phase of an LD cycle. When food is restricted to the light phase, mice will inverse their activity rhythms to coincide with the window of food availability. Food restriction also leads to a clock phase-inversion of multiple tissues: heart, kidney, pancreas and lung57 with

the kinetics of re-alignment being tissue specific78,79. The phase of the SCN clock is unimpacted by the timing of food availability. Food is therefore a primary entrainment cue for peripheral oscillators. Interestingly, arrhythmic Cry1-/-/Cry2-/- double mutant mice also display food entrainable oscillations in hepatic gene expression80. Food driven rhythmic gene expression in circadian mutated mice is generated by the food entrainable oscillator (FEO) that drive food anticipatory activity (FAA) rhythms outside of the circadian clock. Indeed, when the Cry1-/-/Cry2-/- mice were fasted for 24-hours there was a complete loss in their rhythmic gene transcription80.

Peripheral circadian clock function and FEO driven rhythms are separate biological phenomena but both depend upon core circadian clock elements. FAA rhythms are lost in liver specific Per2 mutant mice due to a loss of β-hydroxybutyrate (βOHB) signaling. Adenovirus mediated re- introduction of Per2 to these mutant mice rescued the βOHB phenotype via the additive effects of PER2 and PPARα mediated activation of Ctbp1α which is upstream βOHB production81.

Restricted feeding phase re-alignment of the circadian clock has been argued to involve PPARα and REV-ERBα signaling. PPARα is a metabolically sensitive transcription factor that promotes

Rev-erbα transcription when food is presented at an atypical phase of the behavioral cycle79. This coupled with REV-ERBα stabilization and phosphorylation during food restriction82 likely results in a REV-ERBα induced phase-shift of Bmal1 that alters the timing of BMAL1:CLOCK transcriptional activation on downstream targets. PER2 differentially interacts with the nuclear receptors PPARα and REV-ERBα which may explain the difference between FEO and food restriction circadian realignment. PER2-PPARα serve as activators of Bmal1 via activity at a

Bmal1 regulatory region. PER2-REV-ERBα repress Bmal1 by binding at the Bmal1 promotor region83. PER2 regulation of Bmal1 is therefore context dependent. Under clock mutated conditions PER2-PPARα interaction drive FEO through PPARα nutrient sensing. In this scenario the mutated TTFL renders PER2 as a co-factor for PPARα rather than a core TTFL clock element.

When the clock is functional the interaction between PER2-REV-ERBα may re-align the clock to the new feeding schedule by Bmal1 regulation. Interestingly, Per2 is the only core clock element

that maintains its oscillations within the liver of hepatic specific circadian clock mutants84. Whether this is due to its connection to the FEO, which would still be functional in liver specific clock mutant mice, has yet to be determined, but would offer further validation of the connection between PER2 and FEO rhythms. Circadian and FEO driven rhythmic gene expression in peripheral tissues are influenced by the timing of food intake. Peripheral clock sensitivity is great enough to inverse core clock gene expression in mice with daytime only restricted food access85. This highlights the autonomous nature of peripheral clocks which renders the top down SCN model overly simplistic.

≈50% of protein coding genes are rhythmic somewhere in mice86. Which tissues these genes are oscillating in and what role they have in tissue specific rhythmic biological output is continuously being discovered.

Using liver specific Bmal1 knockout (KO) mice, Lamia and colleagues (2008) showed hepatic Bmal1 is required for rhythmic expression of multiple glucose metabolism genes, including

Glut2. Deletion of Bmal1 resulted in constitutively low GLUT2 that was correlated with attenuated hepatic release of glucose and hypoglycemia77. ClockΔ19 and pancreas-specific Bmal1KO mice are hyperglycemic due to reduced insulin secretion87. Comparison of clock-controlled genes

(CCGs) shared between the liver and pancreas found acetylation at promotor and enhancer regions is tissue-specific88, implicating chromatin structure as an input to tissue specific circadian transcription. Indeed, CLOCK:BMAL1 and the pancreas-specific transcription factor PDX1 bind to insulin secretion factors at similar genetic loci of the pancreas. These data and the lack of shared rhythmic transcripts between tissues50 highlight the tissue specificity of clock function. To gain a greater in vivo understanding of tissue specific clock function would require the generation of more tissue specific clock mutants. Generation of circadian clock mutant mouse lines is however time consuming, expensive and could result in non-circadian phenotypes. Germline

Bmal1KO mice prematurely age, a phenotype that was proposed to be evidence for the need to maintain circadian clock activity to avoid premature aging89. Conditional Bmal1KO mice, however, do not prematurely age. The premature aging in germline Bmal1KO mice is likely the result of

altered activity of Bmal1 during development rather than a circadian relevant phenotype90.

Circadian conditional knockout models alleviate some off-site concerns by offering precise control over when and how long clock ablation occurs. An alternative approach to in vivo evaluation of tissue specific clock function are in vitro 3-dimensional (3D) tissue organoid models. 3D organoids couple enhanced recapitulation of in vivo tissue complexity with the ease of in vitro manipulations.

For circadian studies they can be used to study endogenous tissue clock function outside of SCN input. Further, they present robust circadian clock oscillations91. Previous work leveraged mouse intestinal organoids to characterize the 1:1 and 1:2 coupling of the circadian clock and cell-cycle rate of stem cell and progenitor cells, respectively92. Progress in establishing higher complexity in vitro organoid models will provide an efficient path to extend our knowledge of the role endogenous peripheral oscillators have in driving tissue specific rhythmic biological processes.

1.4 Human intestinal organoids and enteroids as models for extending our

understanding of the endogenous clock of the intestine

The intestine is responsible, among other things, for nutrient uptake and acting as a barrier to the trillions of commensal/pathogenic microorganisms present within its lumen. To maintain its function, the intestinal epithelium continually renews itself. Intestinal cell renewal is driven by resident Lgr5+ intestinal stem cells (ISCs) found in close proximity to Paneth cells (PCs) within the ISC niche at the crypt base93. PCs maintain the ISC niche via release of WNT which regulates

ISC function via R-spondin augmented signaling on the LGR5+ ligand of ISCs94–96. ISC division generates daughter cells that move through the +4 position, out of the ISC niche, to become progenitor cells within the transit-amplifying (TA) zone of the crypt97. Migratory TA cells move up the crypt to the villus region of the intestine and differentiate into one of the multiple intestinal cell types: enterocyte, enteroendocrine, goblet or Paneth cell98. Using a novel R26R-Confetti reporter system, it was shown that all of the intestinal cell types are the direct progeny of ISCs99. Intestinal

crypts maintain their proliferative capacity when grown in vitro and will self-differentiate into intestinal enteroids. To generate enteroids, mouse or human crypt domains are suspended in a

3D basement membrane matrix100,101. Within a day, the crypts will expand to form circular mini- intestines containing a polarized epithelial monolayer surrounding a central lumen with the basolateral surface facing outward. An enteroid will recapitulate in vivo cellular morphology and dynamics outlined above, continually renewing its monolayer with the mature epithelial cell types found in vivo100,102. Human intestinal organoids (HIOs) are a separate 3D in vitro model of the intestine generated from pluripotent stem cells (PSCs). PSCs are differentiated through distinct developmental stages, definitive endoderm (DE) and hindgut tube formation (HG), to generate human intestinal organoids (HIOs); a process mimicking the blastocyst, gastrula, somite and fetal stages of intestinal development, respectively103,104. HIOs serve as a non-invasive method for generating human intestinal tissue from PSCs and have been used to model host-pathogen interactions105,106, cellular development107 and non-pathogenic bacteria interactions108. Initially thought to contain the entire array of mature epithelial cell types103, HIOs have since been found to lack the full spectrum of markers for complete epithelial maturity. The ISC and PC markers,

Lgr5 and Lysozyme, are both present but their secondary markers, Olfm4 and Def5a, respectively, are not expressed in HIOs109. Lacking the complete spectrum of markers for these two primary intestinal cell types underscores the immaturity HIOs. HIOs are also naïve cultures, being grown entirely in vitro without exposure to circulating signals from a basolateral vasculature or an apical microbiome. HIOs can be matured beyond their fetal stage by surgical implantation into the kidney capsule of immune-deficient NOD-SCID IL-2Rγnull (NSG) mice110. While transplanted, HIOs become vascularized by the host mouse circulatory system, which enables them to grow in size, establish a well-defined crypt-villus architecture and express the full spectrum of brush border enzymes and PC markers110. Hierarchical clustering of the transcriptome of non-transplanted vs. transplanted HIOs show non-transplanted samples cluster with in vivo fetal intestinal tissue whereas transplanted samples were matured to an intermediary

state between fetal and adult intestinal tissue109. These in vitro models provide the opportunity to precisely measure intestinal physiology and directly attribute the findings to the intestinal epithelium, not a surrounding tissue type. This unique advantage makes the enteroid/organoid system not only ideal for an epithelial biologist wanting to directly study intestinal epithelial function but also for the chronobiologist wanting to study the function of the endogenous intestinal clock independent of SCN or FEO input.

Due to their multicellular nature, enteroids/organoids can be investigated for intercellular communication pathways. Mouse enteroids maintain robust circadian clock output that equals or outperforms previous homogenous transformed cell lines91. As a population, cell division cycles occur with a periodicity of 12-hours in progenitor cells of the TA zone and 24-hours in ISCs.

Interestingly, neither cell type shows robust circadian clock activity. PCs, however, do have circadian rhythms that promote rhythmic expression of Wnt3A and rhythmic WNT secretion. The rhythmicity in WNT release drives oscillations in the level of WNT within the ISC niche that promote rhythmic cell cycle oscillations of ISCs. Movement of cells out of the ISC niche reduces

WNT mediated cell-cycle regulation, leading to more rapid cell divisions within progenitor cells of the TA region of the crypt92. When intestinal tissue is damaged ISCs must rapidly proliferate to regenerate the damaged epithelial monolayer.

An initial screen of genes involved in the intestinal damage response in Drosophila found

Period (Per), a core clock element, to be indispensable for the proper damage response40. Further analysis found that following damage mitosis occurs rhythmically to replenish the damaged intestinal tissue. This phenotype is lost in Cyc0 and Per0 flies that lack functional clock activity40.

These phenotypes were validated within the mammalian system using radiation induced intestinal damage. Similar to Cyc0 mutant flies, Bmal1-/- arrhythmic mice are incapable of producing rhythmic mitosis to replenish damaged intestinal tissue. Damage-induced intestinal proliferation was mediated by the JNK inflammatory pathway that is downstream of BMAL1 regulation111. The in vivo rhythmic mitotic response, or lack of, were recapitulated in vitro using enteroids isolated

from Bmal1+/+ and Bmal1-/-, respectively. In vivo, the source of the damage induced inflammatory response could be mediated either from the epithelium or non-epithelial myeloid cells. Enteroids were utilized to show that the damage induced signaling response originated from the epithelium.

Further, they provided the opportunity to validate the upstream role of BMAL1 in regulating the

TNF/JNK pathway during the damage response of enteroids111. These studies provide examples of how to uncover molecular clock regulated intestinal physiology using mouse intestinal enteroids. However, detailed roles of the intestinal circadian clock still remain unknown or uninvestigated in human samples. Numerous intestinal circadian-relevant questions could be explored via the human intestinal organoid/enteroid system.

Nutrient uptake85, cellular differentiation112 and the pathogenic response113 have been evaluated for clock input in vivo in mice. Each of these processes has also been investigated using enteroids or organoids106,107,114 but without considering the impact of circadian rhythms.

Human intestinal organoids/enteroids can be used to validate that circadian regulated processes observed in mice are present in human intestinal tissue. Sglt1 and Pept1 are rhythmically expressed in mouse intestinal tissue. When food is limited to the light-phase, anti-phasic to normal food consumption in mice, expression of both transporters re-align to the timing of food availability85. Whether the baseline oscillations and their food restricted re-alignment are a result of intestinal clock function, the food entrainable oscillator or SCN derived signaling is difficult to validate in vivo115. Both transporters are present within enteroids and serve a functional role in moving metabolites across the enteroid monolayer from the basolateral cell culture media into the internal enteroid lumen114. Given the lack of food entrainment or SCN input, enteroids can be utilized to investigate if the endogenous intestinal circadian clock is functionally regulating these nutrient transporters and whether this results in phase-dependent movement of their metabolites.

Human development has been difficult to test in vivo for ethical concerns and in vitro due to poor models. The recent advancement in tissue generation from PSCs has provided a wide range of organotypic model systems that simulate development in vitro. HIOs have been utilized

to elucidate the pathways directing early intestinal development116. Whether they possess circadian rhythms and subsequent rhythmic gene regulation is unknown. Directed differentiation of PSCs into cardiomyocytes activates the circadian clock at the terminal stage of differentiation that results in rhythmic expression of downstream cardiac genes117. Mouse enteroids were utilized to uncover the intricate intercellular PC clock driven regulation of cell cycle in ISCs92. It is compelling to ask whether multicellular HIOs are capable of similar circadian activity seen in PSC- derived cardiomyocytes and to extend this inquiry to whether intercellular communication is involved. If HIOs are arrhythmic, the role of developmental maturation in the onset of clock activity could be tested by transplanting HIOs into the kidney capsule of mice. In vitro kidney capsule matured human intestinal enteroids (kcHIEs) would be isolated to test if HIO epithelial maturation prompts the intestinal clock to become functional. Adult biopsy derived human intestinal enteroids

(bHIEs) would serve as fully matured controls for comparing HIO and kcHIE clock activity against.

The HIO/HIE systems provide an unprecedented opportunity to determine the mechanisms that not only drive intestinal circadian clock function but also their ontogeny.

The influence of the host circadian clock on the pathogenic response was shown in vivo by exposing mice to Salmonella Typhimurium either in the morning or evening113. Subsequent studies using viral118 and protozoan119 pathogens support the pathogenic response is a general circadian regulated process. Rodent models are commonly utilized for host-pathogen studies but vary in their translatability120. In vivo, the response to Salmonella Typhimurium infection varies between mice and humans; humans respond with gastroenteritis and mice have typhoid like symptoms. In vitro, mouse enteroids121 and HIOs106 are both responsive to Salmonella

Typhimurium infection. The mouse enteroid response was characterized by loss of tight junctional integrity121 whereas HIOs responded by upregulating genes involved in the immune response106.

Although not directly compared, the variability in the response of mouse enteroids and HIOs to

Salmonella Typhimurium indicates the in vivo species-dependent variability in the pathogenic response may be recapitulated in vitro. The human enteroid/organoid model system provides an

appealing alternative to in vivo mouse pathogen studies and may offer the greatest translatability to human host-pathogen interactions. Whether multicellular human intestinal enteroid/organoid model systems possess circadian clock activity that is capable of functionally promoting a circadian phase-dependent pathogenic response remains unknown. We utilized toxin B from the intestinal pathogen Clostridium difficile to test for such a response.

1.5 General introduction to Clostridium difficile pathogenesis

1.5.1 Spore formation and transmission

Clostridium difficile, a spore forming gram positive bacterium, is the most common cause of nosocomial acquired diarrhea in the US. In 2011, due to the epidemic strain BI/NAP1/027, there were an estimated 453,000 cases of Clostridium difficile infection (CDI) that resulted in an estimated 29,000 deaths122. C. diff. infection (CDI) manifests following antibiotic treatment123 with diarrhea symptoms that can progress into pseudomembranous colitis and toxic megacolon124.

Spore formation is a primary driver of C. diff virulence. When the spore forming factor, sporulation protein A (Spo0A), is mutated, C. diff can still cause CDI in mice, but transmission becomes attenuated125 because of the inability for spore formation126,127. Mechanistically, spore formation begins after sporulation-associated sensor histidine kinases phosphorylate Spo0A.

Phosphorylated Spo0A then works with the sigma factor SigH128,129 to activate SigF which activates the spore forming signaling cascade via transcriptional initiation130. C. diff spores persist indefinitely and are transmitted to new hosts via the fecal to oral route. Following spore ingestion, it will colonize the gut, lay dormant and maintain an asymptomatic residency until antibiotic treatment. Antibiotic treatment shifts the composition of the commensal microbiota131 and primary/secondary bile acids132,133. Both changes to the intestinal system provide C. diff with a permissive environment to germinate into a vegetative state to begin proliferation and toxin release.

1.5.2 The Clostridium difficile pathogenicity locus

The C. diff genome contains a 19.6kb pathogenicity locus (PaLoc) harboring its toxins,

(Tcd) A, B, C, R and E134. Following germination, C. diff will begin virulent activity via release and action of TcdA and TcdB. The other three toxins within the PaLoc serve regulatory roles over

TcdA and TcdB. TcdR operates as a sigma factor and activates transcription of both TcdA and

TcdB135. TcdC is a repressor of TcdA/TcdB transcription, interfering with TcdR mediated RNA- polymerase recruitment to their promoters136. Correlative evidence supporting the repressive role of TcdC is provided by hypervirulent C. diff strains harboring mutations in the TcdC region of the

PaLoc137. TcdE is responsible for extracellular release of TcdA and TcdB from C. difficile138.

Following TcdE-mediated release of TcdA/TcdB into the intestinal lumen, both TcdA and TcdB will interact with their receptors on intestinal epithelial cells (IECs) to begin their pathogenic breakdown of the IEC monolayer. TcdA receptors on IECs remain elusive, but receptors mediating TcdB cellular uptake have been established.

1.5.3 Toxin B cell receptors

Independent shRNA and CRISPR/Cas9 screens of the HeLa 139,140 and Caco-2 cell lines141 were performed to identify candidate TcdB target receptors. In each case, random mutagenesis was performed to generate thousands of single gene mutated clones. Clones were then repeatedly challenged with TcdB as a selection factor to find which gene mutation provided resistance to TcdB toxicity139–141. Chondroitin sulfate proteoglycan 4 (CSPG4), poliovirus receptor- like 3 (PVRL3) and Frizzled1, 2 and 7 (FZD1, FZD2 and FZD7) were identified as potential TcdB receptors. Knockout (or knockdown) of CSPG4139,140, PVRL3141 and FZD1, 2 and 7140 resulted in the reduction of TcdB induced cytopathic or cytotoxic effect. Systematic evaluation of each of these receptors in mediating TcdB cytopathic effects was carried out within a single study140.

A protection assay was used to evaluate if the extracellular regions of the identified receptors could mitigate the cytopathic effects of TcdB. Protection was assessed by preincubating

TcdB with the extracellular domain of potential TcdB receptors prior to live cell exposure. Inhibition of cell rounding indicates that pre-incubation of TcdB with a receptors extracellular domain prevents TcdB from using that receptor for cell entry into the live cell they are exposed to140.

Protection assays were performed for Caco-2, HT-29 and HeLa cells. CSPG4 extracellular domain (CSPG4-EC) alone attenuated TcdB entry into HeLa cells. Addition of the FZD2 cysteine rich domain (FZD2-CRD), a TcdB binding domain, was required along with the CSPG4-EC to prevent TcdB cytopathic effects in HT-29 cells. TcdB entry into Caco-2 cells was only prevented by the FZD2-CRD. CSPG4 was shown to be expressed at high and moderate levels in HeLa and

HT-29 cells, respectively, and not expressed in Caco-2 cells. CSPG4 expression data therefore assist to explain the protection assay data. CSPG4-EC inhibited TcdB cytopathic effects in HeLa cells that possess high CSPG4 expression. Caco-2 cells, which lack CSPG4, have no protection from TcdB by pre-binding with CSPG4-EC, the FZD2-CRD is needed to prevent TcdB entry into

Caco-2 cells140. PVRL3 did not mediate cellular entry to TcdB in Cspg4-/- HeLa cells or protect

Caco-2 cells when its ecto domain was pre-incubated with TcdB in a protection assay140. This systematic evaluation provided insights uncovering receptor mediated uptake of TcdB in cell lines commonly used to investigate the TcdB response. Cell type-specific immunohistochemistry found

CSPG4 was expressed in the intestinal sub-epithelial myofibroblasts (ISEMF), and FZD2 and

FZD7 were expressed on IECs (Tao et al., 2016). Interestingly, the in vivo data supporting a direct role for CSPG4 in TcdB pathogenesis were gained from intraperitoneal (IP) injection of TcdB139.

ISEMFs would be more accessible to non-luminal injected TcdB compared to IECs which may have facilitated the IP injection phenotypes139. Follow-up studies are needed to discern whether

CSPG4 is involved in the luminal response to TcdB. Although PVRL3 was not found to be playing a relevant role by Tao and colleagues, it was shown to be involved in TcdB pathogenesis both in vitro and in vivo141. More work will need to be done to remedy these contrasting findings. After

cell receptor binding, TcdA and TcdB are internalized by endocytosis to begin pathogenic breakdown of cell monolayers that results in cell death.

1.5.4 Mechanism of action for TcdA-/TcdB-mediated cell death

Once bound to a cellular receptor, C. diff TcdA and TcdB are internalized via non-clathrin or clathrin-mediated endocytosis142,143. Following low pH exposure within the endosome144,145 and host derived inositolphosphate signaling146,147, the toxin’s cysteine-protease domain (CPD) will cleave the glucosyl transferase domain (GTD), releasing the GTD into the host cell cytosol. Once in the cytosol, the GTD targets Rho family small GTPases (RHOA, RAC1 and CDC42). The GTD utilizes UDP-glucose to glucosylate the small GTPases, rendering them inactive148. The central role of UDP-glucose in supporting toxin pathogenesis was shown by UGP2-/- (the enzyme that generates UDP-glucose) HeLa cells having 3,000-fold resistance to TcdB140. Rho GTPases are responsible for organizing the actin cytoskeleton149. Therefore, the inactivation of Rho GTPases results in cytoskeleton breakdown. In vitro toxin mediated cytoskeleton breakdown manifests with the hallmark cytopathic cell rounding phenotype that eventually leads to cell death. The mechanism that promotes cell death, from either toxin, is still not completely understood and is a subject of active investigation. In vitro studies attempting to understand C. diff toxin-mediated cell death pathways have made conflicting findings. Part of the variability can be attributed to differences in study design such as, 1) the C. diff strain a toxin was isolated from, 2) toxin concentration, 3) length of toxin exposure and 4) in vitro cell line selection. Although these differences make it challenging to translate findings across studies, common themes related to

C. diff toxin mediated cell death initiation have emerged.

Both TcdA and TcdB induce cytopathic cell-rounding in vitro. Following rounding, TcdA induces cytotoxicity via apoptosis, activating the apoptotic markers Caspase-3/7150–152. The kinetics of TcdA apoptotic activation were slow, compared to TcdB (see below), occurring no

earlier than 24-hours post-TcdA exposure. GTPase glucosylation mediated the TcdA-dependent cell death response, because TcdA-deficient in GTD activity showed either reduced or no apoptotic cell death150,151. Cells exposed to low concentrations of TcdB (1-10pM equal to

≈0.27ng/mL) activated Caspase-3/7 and the apoptotic cell death response, similar to TcdA.

However, TcdB-dependent apoptotic cell death began 6-hours post-TcdB exposure instead of at

24-hours as was shown for TcdA152. TcdB added at nM (equal to ≈270ng/mL) concentrations resulted in necrotic cell death, rather than apoptotic. High concentration of WT and GTD-deficient

TcdB induced identical necrotic cell death. These data indicated that GTPase glucosylation is dispensable for TcdB mediated-necrotic cell death activation at high concentrations. Instead of

GTPase inactivation, high concentrations of TcdB promoted necrosis as a result of ROS production via activity of NADPH oxidase153. TcdB has also been connected to autophagic cell death. Using human colonocytes, Chan and colleagues found that TcdB exposure resulted in autophagic flux after 6-hours. Mechanistically, this was explained by TcdB treatment increasing phosphorylation of phosphatase and tensin homologue (PTEN) while simultaneously decreasing mTOR phosphorylation. mTOR is an inhibitor of autophagy that has lower activity when phosphorylated, and PTEN is a negative regulator of mTOR with heightened activity when phosphorylated154. The activation of an upstream regulator of autophagy along with other phenotypic evidence provided the first insight into TcdB cytotoxicity being regulated outside of apoptotic or necrotic pathways. In summary, TcdA has been reported to promote GTPase mediated apoptotic cell death after approximately 24-hours of exposure whereas TcdB promotes

GTPase apoptotic cell death within 6-hours, at low concentrations. In contrast, high concentrations of TcdB activate the necrotic cell death pathway through either NADPH oxidase mediated ROS production or autophagy. In vivo, toxin induced cell death mechanisms result in the breakdown of the IEC barrier. IEC barrier breakdown is advantageous for C. diff because it provides an entry point to the intestinal sub-epithelium for additional nutrient sources. Rodent models have been utilized to understand the in vivo role of either C. diff toxin during pathogenesis.

Two studies utilizing a hamster model to determine the in vivo virulence of single toxin producing Clostridium difficile strains reported conflicting results. The first study showed C. diff associated mortality was comparable between strains containing TcdB (TcdA-B+) only and WT whereas TcdA only strains (TcdA+B-) were not as pathogenic155. The second found TcdA+B-,

TcdA-B+ and WT strains were equally lethal although, TcdA+B- only strains required extended time to induce mortality156. Clinically relevant TcdA-B+157 C. diff strains have been identified and a recent report also described clinically relevant Clostridium difficile strains that produce predominately TcdA and not TcdB158. It thus appears that both TcdA and TcdB are capable of producing CDI related pathogenic effects in a clinical setting.

To uncover strain specific differences in TcdB activity, Lopez-Urena and colleagues attempted to directly correlate in vivo phenotypes with in vitro mechanistic understanding using

TcdB from different C. diff strains. They characterized TcdB pathogenesis in vivo with a mouse ileal loop assay using TcdB from the hypervirulent BI/NAP1/027 (TcdBNAP1) strain and the VPI

159 10463 (TcdBVPI) strain . In vivo, only TcdBNAP1 elicited a response characterized by greater immune activation and histological damage. Follow-up in vitro assays were performed using TcdB from both strains. HeLa and CHO cells were equally sensitive to the cytopathic effects of either

TcdB, and NIH3T3 cells had a greater cytopathic response to TcdBVPI. Differences in cellular entry were shown using a pre-binding assay in HeLa cells with a variant NAP1 (TcdBNAP1v) strain of

TcdB. The TcdBNAP1v is incapable of RHOA glucosylation but has similar cell receptor binding as

WT TcdBNAP1. Pre-binding with TcdBNAP1v caused non-glucosylated levels of RHOA to increase in

159 TcdBNAP1 treated cells. TcdBNAP1v pre-treatment did not attenuate TcdBVPI RHOA glucosylation .

These data indirectly show that TcdBVPI and TcdBNAP1 enter cells through different mechanisms.

Tao and colleagues found differential TcdB receptor presence can account for variability in the

TcdB response across cell lines140. In this study the comparison was made between in vivo and in vitro findings. In vivo TcdBVPI elicited no pathological effects but when added in vitro the toxin was equally virulent as TcdBNAP1. Although not tested, a possible explanation could be that the

receptor for TcdBNAP1 is present both in vitro and in vivo whereas the receptor for TcdBVPI was present only in vitro. Support for this hypothesis was provided by the pre-binding assay that found

TcdBVPI enters cells through a different mechanism than TcdBNAP1. It is interestingly to speculate an alternative hypothesis where TcdBVPI is operating through CSPG4 receptor uptake. If so,

CSPG4 was found only in ISEMFs of in vivo intestinal tissue140 which would not facilitate intraluminal injected TcdBVPI entry into IECs.

The 3D intestinal organoid/enteroid model system has yet to be widely utilized for pathogenic studies. Current in vitro systems used to understand C. diff toxin virulence are limited in their translatability due to poor recapitulation of the intestinal epithelium and cell type dependent variability in toxin receptor presence. Better models are needed to accurately represent the human response to C. diff toxins. In vivo C. diff pathogenic studies cannot be ethically carried out in humans. However, enteroids isolated from human intestinal biopsies can be used to test how human derived samples respond to purified C. diff or its toxins. Human enteroids are a multicellular system that more accurately model in vivo human intestinal tissue compared to previous homogenous intestinal cell lines. Expression of C. diff toxin receptors is therefore likely to be more representative of in vivo human intestinal tissue. Further, 3D cells can be dissociated into a 2D monolayer with proper intestinal polarity. The 3D to 2D transition could allow for studies aimed at understanding apical and basolateral C. diff toxin responses. Another benefit to the 3D intestinal system is the presence of robust circadian rhythms. Time-of-day influence in Clostridium difficile pathogenesis has yet to be evaluated.

The goal of the work presented in the following chapters was two-fold. First, we sought to characterize the ontogeny of the intestinal circadian clock. We leveraged the HIO model system to characterize circadian clock activity during early intestinal development. Previous reports show differentiation of PSCs activates the circadian clock. We hypothesized that in vitro differentiation of induced PSCs (iPSCs) into HIOs would similarly activate the circadian clock. We extended the

HIO analysis by determining clock output within kidney capsule matured and patient biopsy HIEs.

The circadian clock has yet to be investigated in any human multicellular intestinal model.

Therefore, we have performed the first comprehensive evaluation of the circadian clock across

3D multicellular intestinal samples that are representative of multiple human developmental stages. The second goal was to evaluate the functional role of the clock in promoting a circadian phase-dependent response to Toxin B (TcdB) from the intestinal pathogen Clostridium difficile.

Time-of-day variability in the pathogen response has been previously established using multiple microbial pathogens. It has yet to be shown whether an isolated bacterial toxin is capable of eliciting a circadian phase-dependent response. We hypothesized that human intestinal samples that possess circadian clock activity would respond to TcdB, a primary virulence factor of C. diff, in a circadian phase-dependent manner. Our data show that the circadian clock is absent from the PSC to HIO stage of in vitro differentiation. In contrast, circadian oscillations are observed in kcHIEs. kcHIE clock output is similar to clock function in bHIEs but is not equivalent. The mechanism that activates clock function in transplanted HIOs were not determined but could be the result of in vivo circulating factors, such as insulin160, that entrain peripheral circadian clocks.

Future work should be dedicated to testing in vitro manipulations that elicit HIO clock activity to provide direct evidence for what is responsible for the onset of circadian rhythms during HIO transplantation. We have also found that the necrotic cell death response to C. diff toxin B is circadian phase-dependent in samples with a functional clock. Interestingly, the necrotic cell death response to TcdB was anti-phasic between mouse and human enteroids. These findings correlated with out-of-phase expression of core clock genes and anti-phasic expression of Rac1 in mouse and human enteroids. More effort is still needed to validate the primary role of Rac1 in driving the circadian phase-dependent response to C. diff toxin B and to understand how the circadian clock might influence the intestinal response to TcdA. In this thesis we provide foundational insights on circadian clock activity in novel 3D in vitro samples and show they functionally regulate the pathogenic response to a common intestinal pathogen.

2. Ontogeny of the intestinal circadian clock

2.1 Abstract

Circadian clocks are responsible for organizing biological processes to occur in coordination with light-dark cycles and food intake rhythms. The developmental timepoint when the molecular circadian clock begins rhythmic activity in humans is not well understood.

Pluripotent stem cell (PSC) derived human intestinal organoids (HIOs), kidney capsule matured human intestinal enteroids (kcHIEs) and patient biopsy derived human intestinal enteroids

(bHIEs) were used to characterize clock activity at multiple developmental timepoints. PSC differentiation into HIOs was used as an in vitro mimic for human fetal intestinal development.

Matured fetal intestinal tissue was modeled with kcHIEs and bHIEs represented fully developed adult intestinal samples. PSC cultures did not possess oscillations. Circadian clock activity was observed in terminally differentiated HIOs but the oscillations were not robust, characterized by low amplitude, unsustainable rhythmic output. kcHIEs and bHIEs possessed equally robust clock activity as measured by Bmal1-luc activity. Amplitude of core clock gene expression was however lower in kcHIEs compared to bHIEs, indicating kidney capsule maturation activates clock activity to an intermediate state between in vitro PSC derived HIOs and bHIEs. We conclude circadian clock activity is not present in an in vitro model of human intestinal development. In vivo maturation was successful in promoting clock activity in kcHIEs but not to an equivalent state as fully matured bHIEs. We have thus modeled circadian clock ontogeny in the human intestine using

3D in vitro human intestinal samples spanning multiple developmental timepoints. These data provide a foundation for future studies attempting to understand what prompts the intestinal circadian clock to become active and provide insight on the developmental timepoint when circadian regulated intestinal biological processes begin their rhythmic function.

2.2 Introduction

The SCN coordinates circadian clocks in peripheral organs to external light-dark cycles.

During pregnancy, the SCN extends its role to coordinate perinatal circadian rhythms in the developing offspring. When pregnant rats were injected with 14C deoxyglucose either during the subjective day or evening, they responded with greater 14C uptake within their brains during the subjective day161. Enhanced uptake during the subjective day was maintained following LD cycle inversion, highlighting correlation between the observed phenotypes and the timing of light exposure. Deoxyglucose uptake was also time-of-day (TOD) dependent in the brains of developing fetuses of the mother; higher during the subjective day, phase-aligned with the maternal uptake161. SCN lesion prompts a complete loss of circadian behavioral and endocrine activity in adult rodents162,163. Maternal SCN lesion at day 7 (D7) post-gestation was therefore used to ablate the circadian system to confirm its influence in driving the TOD dependent deoxyglucose brain uptake in developing fetuses. Daytime deoxyglucose uptake was lost in fetuses within SCN-lesioned rats but not in non-treated or sham control fetuses164. Evidence for maternal circadian input to post-birth neonate pups was provided by determining rhythmic pineal

N-acetyltransferase (NAT) activity, an early indicator of rat circadian activity post-birth165. NAT was used to assess circadian activity in 10-day old neonate rats born from control or SCN lesioned mothers. Control neonates had high NAT during the evening that fell to a trough during the day whereas neonates born from SCN lesioned rats had constitutive NAT activity throughout the day.

Importantly, neonates born from SCN lesioned mothers were capable of rhythmic drinking behavior at D30 post-birth164. These early data showed that circadian signals originating from the maternal SCN directly impact a developing fetus and newborn neonate pups. The offspring born from circadian disrupted mothers can however eventually establish rhythmic behavioral activity of their own following further development. The above studies focused on the brain and SCN lesions because it was assumed the SCN was the only maternal tissue capable of relaying circadian signals to the developing fetus. Further investigations showed that clocks are present throughout the body166 and can operate outside of direct SCN input167. These insights generated a hypothesis

that maternal signaling from non-SCN sources could drive the behavioral rhythms found at D30 post-birth. To address this hypothesis, whole body arrhythmic mPer1Brdm1/mPer2Brdm1 and mPer2Brdm1/Cry1-/- double knockout mothers were used to remove maternal SCN and peripheral clock input to the developing fetus168. By crossing the double mutant females with WT males, the resulting heterozygous offspring were circadian competent but lacked maternal circadian input.

Wheel running activity was monitored for each of the heterozygous offspring and found to be the same as double WT parental controls168. A significant weakness in this design however was the co-housing of new-born neonates with both the arrhythmic mother and WT father. Activity patterns from either parent could influence the recording of behavioral activity or the activity of the neonates themselves. The heterozygotes were single-housed at 7-weeks of age and maintained their circadian behavior168. The transgenic mouse work successfully showed, when tested for at a later timepoint, that offspring of whole-body circadian arrhythmic mothers still possess behavioral rhythms indicating the presence of a functional clock. Loss in maternal circadian signaling impacts the postnatal time of circadian clock onset, not the ability to establish a functional clock itself. Molecular mechanisms regulating maternal signal-mediated synchronization of the clock during the perinatal period of development remain unknown.

Peripheral clocks are coordinated by neuronal and humoral factors from the SCN74,75 and food intake rhythms57, investigations have sought to uncover similar maternal-mediated signaling mechanisms regulating the circadian clock during early development.

LD cycles are a primary driver of circadian systems. Humans are diurnal organisms that preferentially sleep at night following evening melatonin release. The tight coupling between the circadian clock and rhythmic melatonin release has resulted in dim light melatonin onset (DMLO) being the current test of choice for clinical evaluation of an individual’s circadian phase169.

Melatonin was therefore hypothesized to be playing a role in SCN-mediated signaling from mother to fetus during pregnancy. Constant light exposure is an effective treatment for reducing melatonin release170 and disrupting circadian behavioral rhythms171. Constant light exposure to pregnant

females induces circadian arrhythmicity and a reduction in melatonin release that cannot be rescued by repeated melatonin dosing172. The repeated dosing was however successful in promoting circadian variation in arginine vasopressin (AVP), a component of SCN rhythmic function173, within the SCN of postnatal day 1 (P1) neonate pups born from constant light exposed mothers. Melatonin release is regulated from the pineal gland. Pinealectomized mothers, unable to release melatonin, kept under constant light conditions, similarly gave birth to P1 neonates with rhythmic AVP expression following repeated melatonin dosing. Interestingly, rhythmic AVP production in the SCN did not correlate with rhythmic clock activity in the neonate liver at P1 172.

These data indicate melatonin could be a maternal signaling factor that serves as a circadian zeitgeber to the fetal SCN, but not in peripheral organs.

Nutrient intake is a potent circadian zeitgeber that acts on peripheral clocks in adult mammals57,160. Utilizing a novel Rev-erbα-luciferase (Rev-luc) transgenic mouse line Canaple and colleagues showed that the timing of food intake played a role in clock synchronization in utero174.

By crossing male Rev-luc mice with WT females any luciferase signal present in abdomens of pregnant female mice was indicative of Rev-erbα expression within the fetus. Using this approach, it was shown that Rev-luc signal began oscillating at E18.5 in utero. The Rev-luc signal persisted post-birth and continued to have signal and amplitude amplification throughout postnatal development. Food was restricted to the light phase for mothers either during gestation or immediately after birth to test how maternal food restriction impacts Rev-luc rhythms in newborn pups. Regardless of when food was restricted, it resulted in a phase-inversion of the Rev-luc signal at P20 that was more pronounced when food was restricted following birth174. In this work,

Rev-luc signal was not limited to one organ but assessed as a whole-body signal and argued to be indicative of peripheral clock activity174. These data, along with the melatonin study, suggest during development the SCN and peripheral clocks may be influenced by different maternal derived circadian signals. The precise timing of when core clock genes begin oscillating is also different between the SCN and peripheral clocks. Per1, Per2 and Bmal1 begin oscillating at

postnatal day 3 (P3) in the rat SCN. Oscillations of just these three core clock factors was maintained at P10 with larger amplitude175. Core clock gene expression during development has also been evaluated for both the liver and colon. At E20 Rev-erbα oscillates in both tissues, Cry1 and Per1/Per2 are rhythmically expressed in either the liver176 or colon177, respectively. At P2,

Per2 and Cry1 oscillate in the colon177 and Rev-erbα and Bmal1 in the liver176. Per1 and Rev-erbα were rhythmic in both the colon and liver and Bmal1 only in the colon at P10. At P20 all four genes are rhythmically expressed in both the liver and colon176,177. Prenatal expression of clock genes is limited to repressive elements (i.e. Per1, Per2, Rev-erbα, Cry1). Bmal1 oscillations begin during

P2 and P3 in the liver176 and SCN175, respectively. It is intriguing to speculate if these observations support the hypothesis that Bmal1 serves as a developmental gene, rather than a circadian transcriptional activator, during early development; a notion supported by variability in age-related phenotypes in germline89 and conditional90 Bmal1 knockout mice.

Germline Bmal1 knockout mice have a wide array of early aging phenotypes which were attributed to clock disruption causing early aging89. Follow-up work, however, found that post-birth conditional Bmal1 knockout prevented these phenotypes almost entirely90. The difference in germline and conditional, post-birth, knockout phenotypes serve as evidence for the indispensable role of Bmal1 during development. Under this paradigm, prenatal non-rhythmic

Bmal1 expression could be the result of a currently uncharacterized developmental function.

During the post-birth period, the clock related function of Bmal1 may become activated in response to direct stimulation by environmental zeitgebers, prompting rhythmic Bmal1 expression. Bmal1 is a core transcriptional activator of the TTFL. The lack of rhythmic Bmal1 expression during fetal development suggests that the circadian clock may not have robust activity at the fetal stage of development. The onset of Bmal1 oscillations at P2/P3 indicate that this is the earliest developmental timepoint for TTFL driven molecular rhythms. Finally, the lack of rhythmic Bmal1 in the colon177 at P2 suggests that not all tissues begin their circadian output during the early postnatal period. From P3 to P10 the amplitude of rhythmic genes increases,

culminating with nearly all core clock genes being rhythmic by P20176,177. Rodent development occurs at a rapid pace compared to humans. Gestation lasts only a month in rodents vs. 9 months in humans. Rodents still require tissue and organ development postnatally whereas humans are birthed with most tissues fully formed. Developmental studies utilizing rodents are beneficial in gaining a basic knowledge of the process, but have limitations addressing species-dependent differences. Studies of human development are, however, limited due to ethical concerns.

Fortunately, recent progress in developmental biology and tissue engineering have made significant advancements in generating complex in vitro tissues178. The differentiation process from stem cell to multicellular, 3D tissue in vitro mimics cell differentiation and lineage specification in vivo. These tools can be harnessed to understand how cell fate decisions and differentiation may be impacting circadian clock ontogeny in a tissue specific manner. The lack of SCN input into these in vitro systems offer an additional benefit of modeling tissue specific, endogenous, circadian clock development.

Pluripotent Stem Cells (PSCs) do not possess circadian rhythms and show low expression of multiple core clock genes179. Retinoic acid-induced differentiation of mouse embryonic stem cells activates the circadian clock, whereas driving differentiated cells back into a pluripotent state abolish circadian rhythms180. Sequestered PER1/PER2 to the cytoplasm181 and posttranscriptional regulation of Clock112 have been implicated as regulatory mechanisms that inhibit the generation of circadian rhythms during the early stages of differentiation. Transcriptional profiling of PSC derived cardiomyocytes found that terminally differentiated cardiomyocytes possess functional circadian rhythms that regulate circadian phase-dependent responses to doxorubicin-mediated stress117. These data show that differentiation and the onset of circadian clock activity are coupled to one another in vitro. In each study, the terminally differentiated cell type was homogenous and lacked in vivo complexity. Further, the liver176 and colon177 have differential onset of circadian activity in developing rats, indicating the developmental timepoint when the circadian clock becomes active is tissue specific. There is a knowledge gap regarding

how the above data translate to differentiation induced multicellular samples and other, non- cardiac117 and neuronal180 tissues. We utilized the HIO system to investigate the development of circadian rhythms in a multicellular in vitro model of the intestine. HIOs were grown entirely in vitro from directed differentiation of pluripotent stem cells, devoid of any maternal derived signaling.

HIOs were then transplanted into the kidney capsule of NSG mice to mature beyond their in vitro early fetal like state. Robust circadian rhythms were not present at any stage of the in vitro HIO generation process. Clock output was similar, but not equivalent, in human intestinal enteroids established from transplanted HIOs and human patient biopsies (bHIEs). These findings indicate that unlike human PSC derived cardiomyocytes, in vivo exposure, not in vitro differentiation, is required to activate robust circadian clock activity in multicellular 3D human intestinal samples.

2.3 Methods

2.3.1 HIO generation

HIOs were generated following previously established protocols103 using multiple pluripotent stem cell (PSC) lines, both induced and embryonic, to prevent cell line-dependent biasing. Briefly, PSCs were differentiated through the definitive endoderm stage via the addition of Activin A (Cell Guidance Systems) between day-0 (D0) and D2, BMP4 (R&D Systems) at D1 and CHIR99021 (Stemgent) / FGF4 (R&D Systems) from D3 to D6. The addition of

CHIR99021/FGF4 prompted the formation of 3D hindgut tube spheroids from the definitive endoderm monolayer formed by BMP4 and Activin A supplementation. At D7 hindgut tube spheroids were collected via manual picking with a cut pipette tip and suspended in a 50µL

Matrigel basement membrane matrix dome (Corning) at a density of ≈100 spheroids per bubble.

From D7 to D21 spheroids were grown in gut media: DMEM/F12 (Gibco) supplemented with N-2

(Gibco), B-27 (Gibco), HEPES solution (Millipore-Sigma), recombinant murine epidermal growth factor (Peprotech), L-Glutamine (Gibco), Penicillin/Streptomycin (Gibco). At D21 sample density

was reduced by re-suspending differentiated spheroids at a density of ≈1-3 per 50µL Matrigel bubble. Samples were grown for an additional 14-days in gut media, resulting in terminal HIO cultures103. Individual HIOs were considered biological replicates for all experiments.

2.3.2 Generation and maintenance of kidney capsule-matured- and biopsy derived-

human intestinal enteroids

HIOs, following the 1-month differentiation protocol, were transplanted into the kidney capsule of immune-deficient NOD-SCID IL-2Rγnull (NSG) (Jackson Laboratories) mice for maturation, as previously described110. Crypt domains from transplanted HIOs were isolated three months after transplantation and suspended in Matrigel domes to self-differentiate into mouse kidney capsule-matured human intestinal enteroids (kcHIEs)101. Frozen duodenal patient biopsy derived human intestinal enteroids (bHIEs) were provided by the lab of Dr. Noah Shroyer (Baylor

College of Medicine), thawed and expanded in vitro. Both bHIEs and kcHIEs were cultured following identical protocols. Enteroids were expanded by suspending pelleted enteroids in

Matrigel and plating as three separate 10µL domes/well in a 24-well plate. Intesticult organoid growth medium (StemCell Technologies) was used for enteroid expansion. Enteroids were passaged every 7-10 days via syringe (BD) mechanical disruption. Human enteroid differentiation was initiated by replacing expansion media with differentiation media 2-days post-passaging.

Differentiation media consisted of Intesticult Component A (StemCell Technologies) mixed at a

1:1 ratio with DMEM/F12 supplemented with 15mM HEPES. Human enteroids were maintained in differentiation media at least four days before being used for experiments to allow sufficient differentiation, as indicated by a thick columnar epithelium. kcHIEs derived from separate transplanted mice were counted as biological replicates. bHIEs isolated from separate human patient biopsies were counted as biological replicates. bHIEs were isolated from both male and

female patients spanning a range of ages (24- to 74-years of age) and ethnicities to prevent human sample biasing.

2.3.3 Bmal1-luciferase lentiviral transduction and monitoring of Bmal1 activity

All human samples from PSCs to HIOs, kcHIEs and bHIEs were transduced with the same pABpuro-BluF (Addgene) plasmid DNA. The pAB-puro-BluF plasmid contains 1kb of the mouse

Bmal1 promoter region fused to the luciferase coding region, generating a Bmal1-luciferase

(Bmal1-luc) clock reporter construct182. Plasmid DNA was packaged into lentiviral vectors by the

Viral Vector Core at Cincinnati Children’s Hospital Medical Center (CCHMC). PSC transduction was carried out following a previously published lentiviral transduction protocol92. Transduced

PSCs were Puromycin selected (2µg/mL) (Invivogen) and differentiated into definitive endoderm

(DE) and hindgut tube (HG) cultures, generating PSC, DE and HG clock reporting samples.

Hindgut tube samples were re-transduced following the same protocol used for PSC transduction and differentiated into HIOs to generate Bmal1-luc reporting HIOs. The bHIE and kcHIE transduction protocol was designed by merging previously reported enteroid digestion183 and transduction184 protocols. One high density well of enteroids was collected in a 1.5mL Eppendorf and pelleted three times, serially, to fully remove Matrigel. Cells were digested to single cells via treatment with 400µL 0.5mM EDTA/0.05% Trypsin solution (Gibco) at 37°C for 5 minutes. Trypsin was deactivated with 900µL DMEM F12 (Gibco) supplemented with 10% Fetal Bovine Serum

(FBS) (Gibco). Cell clumps were further digested with repeated pipetting using a p1000 pipette.

Single cells were pelleted and resuspended in viral media containing: 50µL Lentiviral suspension

(titer ≈106-7) (CCHMC), Polybrene 8µg/mL (Millipore-Sigma), CHIR99021 (Cayman), 10µM,

ROCK inhibitor (Millipore-Sigma) and Intesticult Component A/B (1:1 mix) to bring the volume to

400µL. Viral media suspensions were added to one well of a 24-well cell culture dish and spinoculated for 1-hour x600g at 32°C. The plate was then placed in a 37°C, 5% CO2 cell culture

incubator for 6-hours. Following incubation, cell suspensions were added to a 1.5mL Eppendorf, pelleted, suspended in 30µL of Matrigel and plated as three 10µL domes/well in a new 24-well plate. 350µL of Intesticult Component A/B (1:1) media supplemented with 10µM ROCK inhibitor and 10µM CHIR99021 was added to the culture. After 24-hours, media was replaced with 350µL

Intesticult Component A/B (1:1) supplemented with 10µM ROCK inhibitor. On day two post- infection media was transitioned to Intesticult Component A/B (1:1) supplemented with 2µg/mL

Puromycin. Puromycin selection was maintained for 2-weeks before starting experiments.

Bmal1-luc transduced samples were monitored for clock output over 4-days with a

KronosDIO luminometer (ATTO). PSCs were added to the KronosDIO device following two weeks of selection post-transduction. DE generated from transduced PSCs were similarly added to the device to record output over 4-days. DE samples at day-4 were in a pre-HG state. The Pre-HG samples were taken out of the KronosDIO, synchronized, provided new media to promote further

HG differentiation, and placed back into the device to begin recording. Transduced HIOs were plated individually to 35mm dishes for bioluminescence recording. kcHIEs and bHIEs were differentiated for 4-days prior to bioluminescence recording. On day 4, samples were re-plated to

35mm dishes, synchronized and added to the KronosDIO to begin recording. All samples were provided a 1-hour clock synchronization stimulus with 100nM Dexamethasone (Millipore-Sigma) and given new non-Dexamethasone containing media prior to recording. 200µM Luciferin

(GoldBio) was added to each culture for bioluminescence detection.

2.3.4 Timecourse sample collection

Rhythmic gene expression was determined via timecourse sample collection and subsequent real-time quantitative reverse transcriptase polymerase chain reaction (qRT-PCR) analysis. Timecourses were collected under similar protocols, with subtle variations. kcHIEs and bHIEs, 2-days post-differentiation onset, were plated in 7 separate, 12µL Matrigel domes per well

Figure 2.1 – Robust circadian clock activity is not present throughout PSC differentiation into HIOs. Bmal1-luc circadian clock reporter activity in: Pluripotent Stem Cell (PSC - purple), Definitive Endoderm (DE - green), Hindgut Tube (HG - orange) and Human Intestinal Organoid (HIO – grey) cultures A) Raw (top) and detrended (bottom) Bmal1-luc bioluminescence signal during the HIO generation process. Robust circadian oscillations were not observed at any stage of PSC differentiation into HIOs (see Figure 3.4 for comparison). B) Period length and amplitude quantification of rhythmic output in A. Continual differentiation promotes period length shortening culminating with circadian oscillations in HIOs. Amplitude is negligible throughout HIO generation and has wide deviation in HIOs. Data represented as Mean ± S.D. of N≥3. replicates per culture condition. Period length and amplitude were quantified using FFT analysis of detrended Bmal1-luc data. Stably transduced Bmal1-luc PSCs were differentiated through the HG stage by changing media, supplemented with luciferin, at the indicated timepoints during the differentiation protocol (section 2.3.1). HG spheroids were again transduced with the Bmal1-luc reporter to monitor HIO rhythms. All cultures were established from iPSC line 252-5. and HIOs were plated in 3, 50µL Matrigel domes. All samples were plated to 6-well plates and had 2mL of either HIE differentiation or HIO media added per well. 24-hours post-plating, enteroids were synchronized with 100nM Dexamethasone for 1-hour. 3mL of media was added to each well post-synchronization to overcome nutrient depletion during the 70-hour collection period. Sample collection occurred between 24- and 68-hours post-synchronization for kcHIEs and bHIEs. Human enteroids were not collected during the first 24-hours post-synchronization to

avoid the inclusion of synchronization-induced, rather than circadian clock-regulated, oscillations in downstream analysis185. HIOs did not show evidence for robust circadian rhythms and had low relative expression of core clock genes. Therefore, HIO samples were collected between 4- and

44-hours post-synchronization. At each collection timepoint, pelleted cells were suspended in

1mL TRI reagent (Molecular Research Center) and snap frozen in liquid N2. Snap-frozen samples were stored at -80°C until all timepoints were collected. Frozen samples were then thawed, vortexed, exposed to 200µL Chloroform, mixed by inversion and centrifuged at 15,000RPM for

15-minutes to separate RNA from DNA/protein. The RNA phase was added to a new 1.5mL

Figure 2.2 – HIOs have low, non-circadian, expression of core clock genes. mRNA expression profiles of Bmal1 (black), Rev-erbα (grey) and Per2 (red) from timecourse sampling of HIOs. Expression values obtained by qRT-PCR using gene specific primer pairs. Solid lines = Mean ± S.D. Dotted lines = individual replicates used to generate mean values. Rhythmic clock gene expression was determined using MetaCycle 2D. Data were derived from independent timecourse sampling using HIOs derived from iPSC line 209 and iPSC line 263-10. 263-10 derived HIOs were used for two separate timecourse collections.

Eppendorf, mixed with 600µL of 70% Ethanol, added to an RNeasy mini-kit column (Qiagen) and had RNA purified following the manufacturer protocol. 200ng of RNA from each timepoint was used to generate cDNA via the Superscript III Reverse transcriptase kit (Invitrogen). FAST SYBR green (Applied Biosystems) and gene specific primer pairs (Table 2.2) were used to quantify gene expression via qRT-PCR on a StepOnePlus (Applied Biosystems) 96-well plate reader with TATA-

Box Binding protein (Tbp) expression used as a housekeeping gene.

2.3.5 Statistical analysis

All data are displayed as mean values of at least three biological replicates with error bars representing standard deviation around the mean. KronosDIO bioluminescent traces were quantified for clock parameters via Fast-Fourier Transformation92. PER2::LUC output (Figure 3.3) is the linear slope of the raw bioluminescence output. qRT-PCR and RNA-seq transcript expression profiles were tested for significant rhythms using the R package MetaCycle 2D186 (p- value cutoff of 0.05). All between group comparisons were made using a two-way analysis of variance (Two-Way ANOVA) with Tukey’s post-hoc analysis. Normality of the residuals was confirmed with Shapiro’s test of residuals at a significance of p<0.05. All statistical tests were performed in R Studio version 1.2.1335.

2.3.6 Tables for reagents and primers

Table 2.1 – Key resources table:

REAGENT or RESOURCE SOURCE IDENTIFIER Biological Samples Human duodenum patient biopsy enteroid GEMS Core – Baylor #103 – female, 24- College of Medicine years old Human duodenum patient biopsy enteroid GEMS Core – Baylor #104 – female, 64- College of Medicine years old Human duodenum patient biopsy enteroid GEMS Core – Baylor #109 – female, 44- College of Medicine years old Human duodenum patient biopsy enteroid GEMS Core – Baylor #110 – female, 60- College of Medicine years old Human duodenum patient biopsy enteroid GEMS Core – Baylor #111 – male, 17- College of Medicine years old Human duodenum patient biopsy enteroid GEMS Core – Baylor #179 – female, 27- College of Medicine years old Human intestinal organoid CCHMC Pluripotent N/A Stem Cell Facility Human intestinal organoid Laboratory of James N/A M. Wells Kidney capsule derived human intestinal enteroid Laboratory of Michael #3 from iPSC 263-10 derived HIO A. Hemrath Kidney capsule derived human intestinal enteroid Laboratory of Michael #4 from iPSC 72.3 derived HIO A. Hemrath Kidney capsule derived human intestinal enteroid Laboratory of Michael #5 from iPSC 72.3 derived HIO A. Hemrath 37

Kidney capsule derived human intestinal enteroid Laboratory of Michael eHIE from H1 hESC derived HIO A. Hemrath Chemicals, Peptides, and Recombinant Proteins Activin A Cell Guidance Cat# GFH6 Systems BMP4 R&D Systems Cat# 314-BP-050 CHIR99021 Stemgent Cat# 04-0004-10 FGF4 R&D Systems Cat# 235-F4 Matrigel basement membrane matrix Corning Cat# 354234 DMEM/F12 Gibco Cat# 12634010 N-2 Supplement (100X) Gibco Cat# 17502048 B-27 Supplement (50X), serum free Gibco Cat# 17504044 HEPES solution Millipore-Sigma Cat# H0887 Recombinant murine EGF Peprotech Cat# 315-09 L-Glutamine Gibco Cat# 25030081 Penicillin-Streptomycin Gibco Ca# 15140122 Puromycin Invivogen Cat# ant-pr-1 Trypsin-EDTA (0.05%), phenol red Gibco Cat# 25300054 Fetal Bovine Serum, heat inactivated Gibco Cat# 10082147 Polybrene Millipore-Sigma Cat# TR-1003-G CHIR99021 Cayman Chemical Cat# 13122 ROCK inhibitor (Y-27632) Millipore-Sigma Cat# SCM075 Dexamethasone Millipore-Sigma Cat# D2915 D-Luciferin GoldBio Cat# LUCK-300 TRI reagent Molecular Research Cat# TR-118 Center 4-OH Tamoxifen Cayman Chemical Cat# 14854 Tris Base Ultrapure US Biological Cat# 77-86-1 Sodium Chloride Fisher scientific Cat# BP358-1 Clostridium difficile toxin B List Laboratories Cat# 155L Critical Commercial Assays Sytox Orange nucleic acid stain Invitrogen Cat# S11368 CellEvent Caspase-3/7 green detection reagent Invitrogen Cat# C10423 Experimental Models: Cell Lines Human induced pluripotent stem cell (iPSC) CCHMC Pluripotent 209 Stem Cell Facility Human induced pluripotent stem cell (iPSC) CCHMC Pluripotent 252-5 Stem Cell Facility Human induced pluripotent stem cell (iPSC) CCHMC Pluripotent 263-10 Stem Cell Facility Human induced pluripotent stem cell (iPSC) CCHMC Pluripotent 72.3 Stem Cell Facility Human embryonic stem cell (hESC) WiCell International WA01 Stem Cell Bank Experimental Models: Organisms/Strains Immune-deficient NOD-SCID IL-2Rγnull The Jackson 005557 Laboratory C57BL/6J The Jackson 000664 Laboratory

PER2::LUCIFERASE Yoo et al., 2004 N/A Bmal1Fx/Fx-EsrCRE Yang et al., 2016 N/A Oligonucleotides Primers to detect intact Bmal1 in Bmal1Fx/Fx- Johnson et al., 2014 N/A EsrCRE enteroids. See table 2.2 Primers to detect exon 4 excised Bmal1 in Johnson et al., 2014 N/A Bmal1Fx/Fx-EsrCRE enteroids. See table 2.2 pABpuro-BluF – Bmal1-luciferase Brown et al., 2005 Addgene #46824 Primers for Bmal1, See table 2.2 This thesis N/A Primers for Per2, See table 2.2 This thesis N/A Primers for Rev-erbα, See table 2.2 This thesis N/A Primers for Tbp. See table 2.2 This thesis N/A Software and Algorithms MetaCycle 2D Wu et al., 2016 – N/A github.com/gangwug/ MetaCycle Fast-Fourier transform analysis code Matsu-ura et al., N/A 2016 – R.script in Appendix Rstudio RStudio Team 2015 1.2.1335 – rstudio.com Leica Application Suire X (LasX) Leica Microsystems – 3.7.0_20979 leica- microsystems.com Kallisto Bray et al., 2016 – N/A pachterlab.github.io/k allisto/ Altanalyze Emig et al., 2010 – N/A github.com/nsalomon is/altanalyze Other Intesticult complete organoid growth medium StemCell Cat# 06010 Technologies Intestinal Component A StemCell Cat# 06011 Technologies Conventional needles BD Cat# BD305167 KronosDIO luminometer ATTO Cat# AB-2550 RNeasy mini-kit column Qiagen Cat# 74106 Superscript III Reverse transcriptase kit Invitrogen Cat# 1808008 FAST SYBR green Master mix Applied Biosystems Cat# 4385612

Table 2.2 – List of primers:

Mouse Forward (5’ → 3’) Reverse (5’ → 3’) Source Self- Bmal1 AGTACGTTTCTCGACACGCAATAG TGTGGTAGATACGCCAAAATAGCT designed Rev- Bugge et GTCTCTCCGTTGGCATGTCT CCAAGTTCATGGCGCTCT erbα al., 2012187 Colbert et Tbp GGTGGCAGCATGAAGTGACA GCACAGAGCAAGCAACTCACA al., 2009188

Intact Johnson et AATCACCTTTTGGGGAGGAC CCCTGAACATGGGAAAGAGA Bmal1 al., 2014189 Exon 4 Johnson et excised AATCACCTTTTGGGGAGGAC TCATCAGAGGAACCAGGGTAA al., 2014189 Bmal1

Human Forward (5’ → 3’) Reverse (5’ → 3’) Source Self- Bmal1 ATAGGCCGAATGATTGCTGAG GGAGGCGTACTCGTGATGTTC designed Self- Per2 TACGCTGGCCACCTTGAAGTA CACATCGTGAGGCGCCAGGA designed Rev- Self- CTTCCGTGACCTTTCTCAGCA GGTGCGGCTTAGGAACATCAC erbα designed Self- Tbp TGTGCACAGGAGCCAAGAGT ATTTTCTTGCTGCCAGTCTGG designed

2.4 Results

2.4.1 Bmal1-luc rhythms are not robust in HIOs

Bmal1-luc activity was recorded throughout differentiation with a KronosDIO luminometer

(ATTO). In agreement with previous reports112, iPSCs lacked detectable circadian clock activity.

Oscillations with period lengths of 50- and 34-hours were present at the DE and HG stages of the differentiation protocol, respectively. Due to the short sampling window, we cannot validate whether DE oscillations are sustainable. HIOs oscillated with a period length of ≈24-hours. These data show that differentiation of iPSCs into multicellular HIOs promotes period length shortening, culminating with circadian oscillations in HIOs (Figure 2.1). However, rhythmic amplitude and bioluminescence signal at each stage were not robust (see Figure 2.4 for comparison). HG and

HIO oscillations were also not sustained beyond 50-hours post synchronization (Figure 2.1). To validate the lack of robust clock activity in HIOs, we also performed timecourse sample collection to determine core clock gene expression via real-time quantitative reverse transcriptase polymerase chain reaction (qRT-PCR).

The circadian TTFL generates out-of-phase oscillations between activating (Bmal1) and repressive (Per2, Rev-erbα) elements. To test for phase relationships between core clock elements, HIOs were collected every 4-hours from 4- to 44-hours post-synchronization, two

Figure 2.3 – Overview of the source and method to establish in vitro 3D intestinal organoids and enteroids. Directed differentiation of pluripotent stem cells (PSCs) is used to generate fetal like, terminally differentiated human intestinal organoids (HIOs) (grey). Embedding HIOs into the kidney capsule of mice over a 6-week period prompts their maturation beyond the in vitro terminal HIO state. Human intestinal crypts were isolated from the matured HIOs and suspended in a 3D Matrigel dome to self-differentiate into kidney capsule matured human intestinal enteroids (kcHIEs) (red). Biopsy derived human (blue) or mouse (yellow) intestinal tissue can be digested to isolate crypt domains. Isolated crypts were suspended in 3D Matrigel domes to self-differentiate into biopsy derived human intestinal enteroids (bHIEs) or mouse enteroids. complete circadian cycles. Gene expression was determined via qRT-PCR with clock gene expression normalized to TATA-Box binding protein (Tbp). Canonical core clock genes, Bmal1,

Per2 and Rev-erbα had low expression and arrhythmic profiles in HIOs (Figure 2.2), consistent with the low Bmal1-luc reporter signal (Figure 2.1). We speculate that qRT-PCR is not sensitive enough to detect low amplitude oscillations, which were observed using the Bmal1-luc reporter.

These data show that robust circadian clock oscillations are not present at any stage of the HIO generation process. Although HIOs contain multiple intestinal epithelial cell types103 they

Figure 2.4 – HIO transplantation activates the circadian clock. Bmal1-luc circadian clock reporter activity in kidney capsule matured human intestinal enteroids (kcHIEs – red) and biopsy derived human intestinal enteroids (bHIEs – blue). A) Raw (top) and detrended (bottom) Bmal1-luc bioluminescence signal during the HIO generation process and in Bmal1-luc transduced kcHIEs and bHIEs. Clock output is noticeably more robust in kcHIEs and bHIEs when compared to Bmal1-luc signal throughout the HIO generation process. B) Quantification of period length and amplitude from detrended data in A. kcHIE and bHIE cultures possess circadian rhythms with a period length of ≈24-hours and similar amplitude. PSC through HIO data were re-graphed from Figure 2.1 for comparison. Data represent Mean ± S.D. of N≥3. replicates per culture condition. Period length and amplitude were quantified using FFT analysis of detrended Bmal1-luc data. Between group comparisons in B were made using a Two-way ANOVA with Tukey’s post-hoc analysis. iPSC line 252-5 was used for HIO generation. kcHIE data was generated using kcHIE lines #3, #4, #5 and eHIE. bHIE data was generated from patient biopsies #103, #104, #109 and #179. lack the full spectrum of epithelial cell markers, which reflect their immature, fetal-like state. The stem and Paneth cell markers Lgr5 and Lysozyme, respectively, are expressed equivalently in

HIOs and human tissue109. The expression of secondary stem cell (Olfm4) and Paneth cell (Defa5,

Reg3a) markers are, however, attenuated in HIOs, indicating they do not possess matured intestinal epithelial cell types109. The immaturity of the samples as a whole, or of a specific cell type, may therefore be contributing to the low-level oscillations throughout HIO production (Figure

Figure 2.5 – A fully functional TTFL is not present in kcHIEs. mRNA expression profiles of Bmal1 (black), Rev-erbα (grey) and Per2 (red) in kcHIEs and bHIEs. Samples were collected every 4-hours over two full circadian cycles, 46- hours. Expression values obtained by qRT-PCR using gene specific primers. A) kcHIEs have rhythmic expression of Bmal1 and Rev-erbα. Lack of Per2 rhythms and attenuated expression of Rev-erbα, compared to bHIEs, indicate kcHIEs do not possess a fully functional circadian clock TTFL. B) bHIEs have rhythmic expression of all three clock genes. Rev-erbα and Per2 are properly phase-delayed from Bmal1 by 8- and 12-hours, respectively. Solid lines = Mean ± S.D. Dotted lines = individual replicates used to generate mean values. Rhythmic clock gene expression was determined using MetaCycle 2D. kcHIE timecourse data collected from independent timecourse sampling of line #3 and eHIE. The third kcHIE biological replicate timecourse was generated by combining kcHIEs from lines #4 and #5 for collection at each timepoint. kcHIE line #4 and #5 were both derived from HIOs generated using iPSC line 72.3. bHIE timecourse data was collected from independent timecourse sampling of patient biopsy lines #103, #109 and #179. 2.1) and the attenuated, arrhythmic, expression of core clock genes in HIO samples (Figure 2.2).

Therefore, we transplanted HIOs into the kidney capsule of NSG mice and allowed them to mature over 6-weeks110 to test if HIO maturation could enhance the robustness of circadian activity.

2.4.2 In vivo maturation of HIOs prompts GI circadian clock development

After 6-weeks of HIO maturation in the mouse kidney capsule, human intestinal crypts were isolated from NSG mice to generate in vitro mouse kidney capsule-matured human intestinal enteroids (kcHIEs)101,110. Biopsy-derived human intestinal enteroids (bHIEs) were also established to offer insight on circadian clock activity in a fully matured 3D intestinal enteroid

(Figure 2.3). Both kcHIEs and bHIEs were transduced, via lentivirus, with the Bmal1-luc reporter used previously. In contrast to the HIOs, both kcHIEs and bHIEs showed high amplitude circadian oscillations, confirming in vivo transplantation of HIOs results in the development of robust circadian rhythms (Figure 2.4).

As described above, we performed time course experiments collecting kcHIEs, and bHIEs every 4-hours over two circadian cycles for measurement of core clock gene expression via qRT-

PCR. Both kcHIEs and bHIEs demonstrated circadian rhythms with appropriate phase relationships between positive (Bmal1) and negative (Rev-erbα) clock elements (Figure 2.5).

Interestingly, the expression and amplitude of Per2 and Rev-erbα are significantly lower in the kcHIEs compared to bHIEs, suggesting bHIEs have greater clock activity compared to kcHIEs.

2.5 Conclusions and discussion

The data shown above provide insights into intestinal circadian clock ontogeny by using multiple in vitro models that represent distinct developmental stages. Differentiation of PSCs into

HIOs resulted in a period length shortening phenotype with period lengths of 50-, 34- and 24- hours in DE, HG and HIO cultures, respectively. HIOs possessed low amplitude unsustainable oscillations which was validated by the lack of rhythmic core clock gene expression. None of the three primary clock genes tested had rhythmic output and all had low relative expression. These findings indicate in vitro differentiation of PSCs into HIOs is insufficient to generate meaningful rhythmic output. Differentiation of mouse embryonic, or induced pluripotent, stem cells prompts oscillations of a transduced Bmal1-luciferase reporter construct after 28-days112. Interestingly, a

second study from the same group found that human iPSC’s take longer, requiring up to 90-days of differentiation to initiate circadian oscillations190. Differentiation of human embryonic stem cells

(ESC) into cardiomyocytes however results in rhythmic cultures and transcriptomic oscillations after 30-days117. The lack of rhythmic output in HIOs after a 35-day differentiation protocol align with the extended time to observe clock activity during differentiation of human iPSCs and contrast the 30-day onset of rhythmicity in ESC derived cardiomyocytes. The detailed reason for this discrepancy is currently unknown but could be explained by variability in stem cell line used or tissue specificity in clock function. Tissue specific variability is supported with in vivo rodent data.

Prenatal mouse hearts have been reported to have rhythmic output in up to 4% of their transcriptome when collected between E17-E19112. Colonic clock activation takes a comparatively longer time to become active, not being observed until P20177. The DE, HG, and HIO stages of the HIO differentiation protocol represent mouse E7.5, E8.5 and E16.5 respectively103. Therefore, the lack of robust circadian rhythms in iPSC-derived HIOs and earlier onset of circadian rhythmicity in the ESC-derived cardiomyocytes align with these in vivo rodent data. More work is needed to understand the source of these inter-tissue differences.

In contrast to our findings with HIOs, we observed robust circadian rhythms in kcHIEs and bHIEs. Bmal1-luc activity and Bmal1 gene expression was similar in kcHIEs and bHIEs. However,

Rev-erbα had lower amplitude oscillations and Per2 was not significantly rhythmic in kcHIEs whereas both genes had significant oscillations in bHIEs. These findings support the intermediate maturation state achieved by embedding HIOs into the kidney capsule of mice. Transcriptomic clustering has shown that HIOs and adult intestinal tissue are dissimilar at the genome level.

Transplanted HIOs, compared to these two extremes, frequently occupy an intermediate quantitative space depending on the transcript evaluated. The brush border enzymes Sucrase isomaltase and Trehalase both have low expression in HIOs and high expression in adult intestinal tissue109. Expression of either enzyme was greater in kcHIEs compared with HIOs but not equivalent to adult tissue. Similar findings were made for Paneth cell and Stem cell markers109.

Our data extends upon these findings, showing circadian clock genes are also expressed at an intermediate state in kcHIEs. Having higher expression and more rhythmic output than HIOs but not equivalent to bHIEs.

In this chapter, we utilized a multicellular 3D in vitro system to characterize the ontogeny of the human intestinal clock. Our findings show that in vitro differentiation alone is incapable of inducing robust circadian rhythms in human organoids. Rodent intestinal tissue at a similar development stage as our HIOs also lack rhythmic output of the circadian TTFL. In vivo maturation was successful in inducing clock activity to a similar, but not equal, level as adult biopsy samples.

The detailed mechanisms responsible for circadian clock establishment following in vivo maturation were not evaluated here but could be readily investigated within this system via host environment manipulation (see Section 4.1.2). The intestinal clock is responsible for promoting circadian variation in a wide range of biological processes, including intestinal motility191, nutrient absorption85, innate immunity192 and the pathogenic response113. Rhythmic regulation of these processes is likely to begin at the developmental timepoint when the intestinal circadian clock becomes active. We utilized the above characterization of circadian rhythms within human 3D intestinal samples to provide validity to this premise. All samples were challenged with toxin B from the intestinal pathogen Clostridium difficile at multiple phases of their circadian rhythm.

We hypothesized the presence of a functional circadian clock in 3D intestinal samples would drive a circadian phase-dependent response to Clostridium difficile toxin B (TcdB).

3. The intestinal circadian clock regulates a phase-dependent

response to toxin B from Clostridium difficile

3.1 Abstract

The circadian clock regulates the response to bacterial, viral and protozoan pathogens in mice. Whether similar circadian pathogenic responses are present in humans is difficult to

evaluate due to poor models. Here we utilized mouse and human 3D in vitro intestinal enteroids and organoids to offer a cross-species evaluation of whether the intestinal circadian clock regulates a circadian phase-dependent response to toxin B (TcdB) from the intestinal pathogen

Clostridium difficile. All 3D intestinal samples that possessed clock activity had a circadian phase- dependent necrotic cell death response to TcdB. Interestingly, the phase of greater necrotic cell death was opposite between mouse and human intestinal enteroids. RNA-sequencing analysis show mouse enteroids and patient biopsy human intestinal enteroids (bHIEs) both possess circadian transcriptomes but the phase of core clock genes was species specific. The TcdB target,

Rac1, was also anti-phasic between mouse and human enteroids which correlated with the species-dependent anti-phasic necrotic cell death response to TcdB. Our data highlight unique differences between mouse and human intestinal enteroids, and provide a foundation for human organ- and disease-specific investigation of clock functions using human organoids for potential chronotherapeutic translational applications.

3.2 Introduction

The suprachiasmatic nucleus (SCN) is commonly referred to as the ‘master clock’ of the mammalian system. Located in the hypothalamus, the SCN contains approximately 20,000 neurons that process photic signals to coordinate circadian rhythms in peripheral organs to external light/dark signals72,193. However, peripheral clocks are malleable and can be detached from SCN coordination by non-photic circadian zeitgebers. Mice housed under daytime only restricted feeding protocols realign the phase of clock and metabolic genes in the liver, but not the SCN, to coincide with food availability57,85. Intriguingly, intestinal microbiota composition changes over a circadian cycle194, and can also impact peripheral clock function. Antibiotic induced microbiota depletion results in constitutively high expression of Rev-erbα that abolishes

Figure 3.1 – Bi-directional interactions between the circadian clock, microbiota and immune system. The intestinal microbiota and immune system have rhythmic biological function driven by the circadian clock. However, both the microbiota and immune system can feedback on the clock to impact its function. Dysbiosis and altered immune function have been associated with Clostridium difficile infection (CDI). Based upon the presented interconnectedness of the intestinal clock, microbiota and immune system we believe the circadian clock is likely involved in CDI. rhythmic toll-like receptor (TLR) expression195. Loss of microbiota was also shown to change rhythmic gene expression within the mouse intestinal and hepatic transcriptome while leaving behavioral rhythms intact196. These insights underscore the intimate relationship between intestinal circadian rhythms, microbiota and innate immunity (Figure 3.1). The recent development of tissue-specific conditional Bmal1 rescue mice was instrumental in uncovering intact peripheral circadian rhythms in the skin epidermis and liver in the absence of circadian signaling from the SCN167. Peripheral clocks are therefore more autonomous than previously

thought and, as such, capable of processing input signals from local factors, independent of SCN signaling. To provide functional relevance to endogenous clock function we used human intestinal organoids and enteroids to test the circadian response to Clostridium difficile toxin B. 3D intestinal organoids and enteroids offer the unique opportunity to investigate the role of tissue endogenous circadian input to the pathogenic response in a complex in vitro system that is devoid of SCN or circulating entrainment factors.

Clostridium difficile is a spore forming, gram positive, toxin producing intestinal pathogen.

C. diff. infection (CDI) manifests following antibiotic treatment123 with diarrhea symptoms that can progress into pseudomembranous colitis and toxic megacolon124. CDI pathogenesis includes the luminal release of two toxins, toxin A (TcdA) and B (TcdB), that bind to intestinal cellular receptors139–141 and are brought into the cell via endocytosis143,197. In the cytoplasm, autocatalytic cleavage releases the glucosyltransferase domain of the toxins146 that inactivate the small

GTPases RhoA, Rac1 and Cdc42, resulting in cellular cytoskeleton breakdown, disruption of epithelial barrier function and cell death148,198. In addition, the NOX1 NADPH oxidase complex has also been implicated in promoting toxin-mediated cell death at high concentrations of TcdB153.

Several experimental models have been utilized to investigate C. difficile pathogenesis and each present limitation. In vivo rodent studies have been associated with low reproducibility attributed to confounding variables such as different C. diff. strains, antibiotic regiment to induce susceptibility and cross contamination of animal housing120. The translatability of pathophysiological findings from rodent models to human also cannot be guaranteed. In vitro and ex vivo studies are, comparatively, controllable but fail to recapitulate in vivo cellular complexity or multi-timepoint scalability, respectively. Human intestinal organoids/enteroids offer the opportunity to overcome these limitations by coupling in vitro experimental control with recapitulation of an in vivo human intestinal epithelium. 3D intestinal organoids have been successfully used to gain insights into the pathogenic response to human rotavirus199, E. coli200

and Clostridium difficile toxins A/B105. These studies however did not investigate the role of the circadian clock in the pathogenic response. Circadian mediated time-of-day dependency in the pathogenesis of Salmonella Typhimurium113 and herpes simplex virus118 have been observed in mice. Similar insights into the human pathogenic response are lacking. Therefore, we utilized 3D human intestinal organoids and enteroids to investigate whether the endogenous intestinal circadian clock regulates a circadian phase-dependent response to Clostridium difficile toxin B

(TcdB). We report that the enteroid response to TcdB is circadian phase-dependent in samples with a functional clock. The response to TcdB is anti-phasic between mouse and human enteroids which correlates with out-of-phase clock oscillations and anti-phasic Rac1 oscillations between mouse and human samples. From these data we conclude that the intestinal circadian clock influences the response to TcdB in a species-dependent manner via rhythmic regulation of Rac1.

3.3 Methods

3.3.1 Animals

3-6-month-old C57BL/6J mice were purchased from Jackson Laboratories and sacrificed immediately upon receipt. C57BL/6J mouse enteroids were utilized solely for mouse enteroid timecourse sampling to profile their circadian transcriptome by RNA-seq. Immune deficient NOD-

SCID IL-2Rγnull (NSG) mice were utilized for kidney capsule transplantation of HIOs and were housed at Cincinnati Children’s Hospital and Medical Center (CCHMC). Surgical transplantation of HIOs into NSG was performed following IACUC protocol #2016-0014 (PI: Michael Helmrath).

Mice containing the PER2::LUCIFERASE (PER2::LUC) clock reporting gene have been previously described166. PER2::LUC mice were utilized to provide support for the generation of anti-phasic oscillations following out-of-phase Dexamethasone synchronization and as a control mouse enteroid line for testing the circadian phase-dependent response to TcdB. The Bmal1Fx/Fx-

EsrCRE mouse line90 was established by the lab of Dr. Garret FitzGerald (University of

Pennsylvania) and gifted to us by the lab of Dr. John Hogenesch (CCHMC). Multiple tissues from this mouse become arrhythmic following tamoxifen treatment in vivo90. Enteroids established from this transgenic mouse line were used to validate the need for a functional enteroid clock to produce a circadian phase-dependent response to TcdB. PCR genotyping of enteroids established from Bmal1Fx/Fx-EsrCRE mice was performed utilizing primers listed in Table 2.2. Both the PER2::LUC and Bmal1Fx/Fx-EsrCRE mice were housed in the University of Cincinnati laboratory animal medical service facility, maintained under 14:10 LD cycles and provided food and water ad libitum. Mice were sacrificed via CO2 asphyxiation with cervical dislocation as a secondary method of euthanasia. C57BL/6J, PER2::LUC and Bmal1Fx/Fx-EsrCRE mouse experiments were approved under the University of Cincinnati IACUC protocol #17-01-30-01 (PI:

Christian Hong).

3.3.2 Isolation and maintenance of mouse enteroids

Mouse jejunal enteroids were generated following established protocols201. Briefly, crypt domains from 8cm of mouse mid-small intestinal tissue were isolated. Approximately 150 crypt domains were then added to 40µL of Matrigel (Corning) membrane matrix, suspended as a single

3D Matrigel dome within a 24-well plate and provided 400µL mouse enteroid media. Mouse enteroid media was composed of L-Glutamine (Gibco), Penicillin/Streptomycin (Gibco), N-2

(Gibco), B-27 (Gibco), HEPES solution (Millipore-Sigma), recombinant murine epidermal growth factor (Peprotech), and R-Spondin/Noggin conditioned media generated in-house. Enteroids were passaged every 4-7 days via passaging cells through a syringe (BD). CRE driven Bmal1 exon 4 excision was initiated in vitro by adding 1µM 4-OH tamoxifen (Cayman) for 24-hours to

Bmal1Fx/Fx-EsrCRE enteroids. Bmal1 exon 4 excision was confirmed by PCR 2-days post- treatment, using the Bmal1Fx/Fx primers listed in Table 2.2. Tamoxifen treated Bmal1Fx/Fx-EsrCRE

Figure 3.2 – TcdB phase-response experiment design. Experiment design for testing mouse and human samples for a circadian phase- dependent response to Clostridium difficile toxin B (TcdB). Two sample groups were synchronized either 36- (D36-blue) or 24- (D24-green) hours prior to TcdB addition to generate two out-of-phase control groups. D24- (grey) and D36- (red) TcdB treated samples were imaged at 2-, 24- and 48-hours after TcdB addition to quantify fluorescent cell death dye uptake. enteroids were subjected to two passage/growth cycles prior to experimental utilization. Enteroids were maintained in cell culture incubators set at 37°C and 5% CO2. Enteroids isolated from separate mice were considered biological replicate samples. Enteroids established from the same mouse were utilized as technical replicates. Data generated from technical replicate enteroids were pooled to establish a mean value per treatment group for each biological replicate.

3.3.3 Experiment design to test for a circadian phase-dependent response to TcdB

All samples were density controlled prior to treatment exposure. Density was controlled via manual pipetting under a compound microscope during sample plating to a 24-well plate. 1-2

HIOs were plated in a single Matrigel dome and mouse/human enteroids were plated in groups of ≈15-20 per Matrigel dome. Following plating, two separate circadian phase groups were

generated by adding 100nM dexamethasone (DEX) to one group 36-hours prior to TcdB addition

(D36), and the delayed DEX treatment of the other group by 12-hours, at 24-hours prior to TcdB addition (D24). D36 and D24 samples were then exposed to vehicle buffer (50mM Tris Base

Ultrapure (US biological), 100mM NaCl (Fisher Scientific), pH 7.5) or Clostridium difficile toxin B

(List Laboratories) at either 36- or 24-hours post-DEX treatment, respectively (Figure 3.2).

Vehicle and TcdB treatments were added directly to the cell culture media. TcdB was added at a concentration of 10ng/mL (37pM), 100ng/mL (370pM) and 5µg/mL (18.5nM) to mouse enteroids,

HIOs and human enteroids, respectively. Fluorescent cell death dyes, Sytox Orange (500nM)

(Invitrogen) or CellEvent Caspase 3/7 (2µM) (Invitrogen) were added simultaneously with

Vehicle/TcdB for visualization and quantification of necrotic or apoptotic cell death, respectively.

Brightfield and fluorescent images were acquired at 2-, 24- and 48-hours post-Vehicle/TcdB addition on a Dmi8 inverted microscope (Leica Microsystems). Quantification of fluorescent dye uptake was performed using LasX software (Leica Microsystems). An area of interest for each enteroid tested, at each timepoint post-TcdB, was established by manually tracing the periphery of a sample using the acquired single plane brightfield image (LasX). Average fluorescence intensity within the area of interest was calculated in the associated fluorescent image to account for size variability in intestinal organoids/enteroids. Individual HIOs were considered biological replicates and had data directly pooled to provide treatment averages. Reported fluorescence intensity in each figure is the average of three or more independent biological replicates (as described in section 3.3.2 – mouse and 2.3.1/2.3.2 – human). Mouse and human enteroid values were generated using three or more technical replicates to generate a biological replicate value for each treatment group. All TcdB data were set relative to time- and phase-matched vehicle controls (i.e. fluorescence intensity of D24 TcdB treated samples at 24-hours post- synchronization were set relative to fluorescence intensity of D24 vehicle treated samples at 24- hous post-synchronization) set to 1.

Figure 3.3 – Synchronization of PER2::LUCIFERASE mouse enteroids 12-hours apart drives samples out-of-phase. PER2::LUCFERASE (PER2::LUC) mouse enteroids were synchronized for 1-hour with 100nM Dexamethasone either 12-hours (blue) or immediately (green) before bioluminescent recording. A) Raw (top) or detrended (bottom) bioluminescent recordings of PER2::LUC mouse enteroids synchronized 12-hours apart. Synchronization of separate mouse enteroid cultures 12-hours apart drives their oscillations to be out-of- phase. B) Synchronization timing did not influence quantifiable clock parameters: Period length, Amplitude or PER2::LUC output. Data are shown as Mean ± S.D. of N=3 biological replicates. Period Length and amplitude were quantified by FFT analysis of detrended PER2::LUC data. PER2::LUC output is the linear trendline of raw PER2::LUC bioluminescent data. Comparison’s in B were made using a student’s t-test. Bioluminescence recorded using a KronosDIO luminometer. ≈50 mouse enteroids were plated in a 40µL Matrigel dome within a 35mm dishes and provided 2mL media supplemented with 200µM D-luciferin to record rhythmic luminescence over 4-days. 3.3.4 RNA-sequencing analysis

To perform RNA-sequencing we collected C57BL/6J mouse enteroids and human bHIEs as a timecourse, as outlined in Section 2.3.4. Mouse enteroids were pooled from two separate

Figure 3.4 – Necrotic, not apoptotic, cell death is phase-dependent in PER2::LUC control mouse enteroids. Mouse enteroids were challenged with either 10ng/mL (37pM) TcdB or an equivalent amount of vehicle buffer at two-phases, D24 (green – vehicle, grey – TcdB) or D36 (blue – vehicle, red – TcdB), and monitored for necrotic (Sytox orange – red fluorescence) or apoptotic (CellEvent – green fluorescence) cell death at 2-, 24- and 48-hours post-TcdB treatment. A) PER2::LUC mouse enteroids have greater necrotic cell death in the D36 phase group compared to the D24 phase group at 48-hours post-TcdB addition. B) Apoptotic cell death in mouse enteroids treated with TcdB is greater than vehicle controls at 48-hours post-exposure. The apoptotic response is phase- independent. Data represented as Mean ± S.D. of N≥3 biological replicates generated by pooling data from N=3 technical replicates per biological replicate. All data were normalized to time/phase matched vehicle controls set to 1. *p<0.05, **p<0.01, ***p<0.001. Scale bar = 250µM. mice to mitigate single mouse biasing and bHIEs were from a single patient (#103 – female, 24- years old) biopsy. Samples were collected every two hours for 46-hours starting at 24-hours post

DEX synchronization. Isolated RNA samples (section 2.3.4) were shipped on dry ice for further processing and sequencing analysis (Novogene). Raw RNA-sequencing FASTQ files were directly used to generate expression files using the Kallisto202 embedded algorithm inside the

Altanalyze module203. Kallisto pseudo-aligned each transcript with the Ensemble reference transcriptome (Ensembl version 72) and calculated TPM (transcripts per million), from which

Figure 3.5 – Tamoxifen treated Bmal1Fx/Fx-EsrCRE mouse enteroids lack rhythmic core clock gene expression. Bmal1 (black) and Rev-erbα (grey) gene expression in control PER2::LUC and tamoxifen treated Bmal1Fx/Fx-EsrCRE mouse enteroids. A) Control mouse enteroids possess rhythmic core clock gene expression with out-of-phase profiles. B) Clock gene expression is attenuated and arrhythmic in tamoxifen treated Bmal1Fx/Fx-EseCRE mouse enteroids. Thick lines are Mean ± S.D of N=3 replicates depicted as dotted lines. Rhythmic expression patterns were assigned significance by analyzing biological replicate gene expression profiles using MetaCycle 2D. Altanalyze calculated gene expression. The resulting gene expression summary file was used to run Metacycle 2D (ARS and JTK option)186 with period lengths of 24- to 25-hours and rhythmic genes were filtered using p-value cutoff.

3.4 Results

3.4.1 Intestinal 3D cultures with a functional clock have a circadian phase-dependent

response to C. diff Toxin B

Figure 3.6 – Tamoxifen treated arrhythmic Bmal1Fx/Fx-EsrCRE mouse enteroids do not respond to TcdB in a circadian phase-dependent manner. Mouse enteroids were challenged with 10ng/mL (37pM) TcdB or an equivalent volume of vehicle buffer at two phases, D24 (green – vehicle, grey – TcdB) or D36 (blue – vehicle, red – TcdB), and monitored for necrotic cell death (Sytox orange – red fluorescence) at 2-, 24- and 48-hours post-TcdB treatment. A) Tamoxifen treated arrhythmic Bmal1Fx/Fx-EsrCRE enteroids have a phase- independent increase in necrotic cell death at 48-hours post-TcdB exposure. B) Non-tamoxifen treated Bmal1Fx/Fx-EsrCRE enteroids have greater necrotic cell death in the D36 phase-group vs. D24 at 48-hours post-TcdB addition. Data represented as Mean ± S.D. of N=3 biological replicates generated by pooling data from N=3 technical replicates per biological replicate. All data were normalized to time/phase matched vehicle controls set to 1. **p<0.01, ***p<0.001. Scale bar = 250µM. Circadian clock driven time-of-day variation in the pathogenic response has been established for bacterial113, viral118 and protozoan119 pathogens in mice. A circadian clock dependent response to bacterial toxins in mouse enteroids has not been reported. We therefore tested mouse enteroids for a circadian clock mediated response to purified toxin B (TcdB) from the intestinal pathogen Clostridium difficile. To evaluate if the mouse enteroid response to TcdB is circadian phase-dependent we monitored the TcdB response in separate samples synchronized 12-hours apart, a mimic to morning and evening exposure in vivo. A 1-hour pulse of Dexamethasone (DEX) reproducibly resets the clock within in vitro samples by activation of

clock repressive elements204. Providing DEX to separate PER2::LUC clock reporting mouse enteroid groups 12-hours apart successfully drives their oscillations out-of-phase to one another

(Figure 3.3). Importantly, the out-of-phase cultures were not different in quantifiable clock parameters (Period Length, Amplitude, PER2::LUC output). We therefore tested PER2::LUC samples for a circadian phase-dependent response to either vehicle or TcdB exposure at 36- hours (D36) or 24-hours (D24) post-synchronization. Images were acquired at 2-, 24-, and 48- hours post-TcdB treatment to observe TcdB induced phenotypes (Figure 3.2). Brightfield images indicated that TcdB treatment was effective, driving cytopathic cell rounding in mouse enteroids at 48-hours post-exposure in both phase groups. In vitro TcdB exposure has been observed to promote both necrotic151 and apoptotic cell death105. To determine if TcdB induces either cell death pathway in a circadian phase-dependent manner, we added necrotic (Figure 3.4 - red) and apoptotic (Figure 3.4 - green) fluorescent cell death dyes simultaneously with TcdB. Necrotic dye uptake was significantly greater than control in the D36 TcdB treated phase group at 24-hours post-TcdB treatment. Necrosis in the D24 phase group was not different from vehicle controls at the same timepoint. After 48-hours, both the D36 and D24 phase groups had significantly greater necrotic cell death when compared with vehicle treated controls. Necrotic cell death was however greater in the D36 group compared to D24, indicating TcdB provided at the D36 phase in mouse enteroids results in greater necrotic cell death when compared to the D24 phase. (Figure 3.4 – red). Apoptotic cell death in response to TcdB was greater than vehicle controls only at 48-hours post-exposure. Unlike our findings with necrotic cell death, apoptotic cell death occurred equivalently in the D24 and D36 phase groups (Figure 3.4 – green). Due to this finding, all subsequent analyses tested only for a circadian phase-dependent necrotic cell death response to TcdB, not apoptotic. To validate that the PER2::LUC phase-dependent necrotic cell death response to TcdB is driven by the mouse enteroid clock, we repeated the experiment using enteroids established from the tamoxifen inducible Bmal1Fx/Fx-EsrCRE mouse90.

In vivo tamoxifen treatment of Bmal1Fx/Fx-EsrCRE mice results in the loss of clock activity in multiple tissues90. Tamoxifen treatment of Bmal1Fx/Fx-EsrCRE mouse enteroids recapitulates these phenotypes in vitro as shown by the attenuated and arrhythmic expression of Bmal1 and

Rev-erbα (Figure 3.5). Tamoxifen treated Bmal1Fx/Fx-EsrCRE enteroids responded to TcdB with enhanced necrotic cell death at 48-hours post-TcdB, similar to PER2::LUC mouse enteroids.

However, cell death was no longer phase- dependent, being activated similarly in the D24 and

D36 TcdB treated groups (Figure 3.6 (A)). Non-tamoxifen treated Bmal1Fx/Fx-EsrCRE enteroids possessed a phase-dependent response to TcdB that is identical to the PER2::LUC enteroid response, greater in the D36 phase compared to the D24 (Figure 3.6 (B)). These data show the mouse enteroid phase-dependent response to TcdB is circadian driven, requiring the presence of a functional clock to be observed. In chapter 2 we characterized clock activity across multiple human 3D intestinal samples. We therefore sought to validate our findings in mouse enteroids with human derived samples which also provides translatable insights into the role of the endogenous human intestinal clock during the response to TcdB.

HIOs, kcHIEs and bHIEs were tested for circadian phase-dependent responses following the same protocol used for mouse enteroids (Figure 3.2). HIOs lacked a phase-dependent

Figure 3.7 – Necrotic cell death is phase-dependent in human 3D intestinal models with a functional clock. TcdB was added at 100ng/mL (0.37nM), 5,000ng/mL (18.5nM) and 5,000ng/mL (18.5nM) to HIOs, kcHIEs and bHIEs, respectively, at two phases D24 (green – vehicle, grey – TcdB) or D36 (blue – vehicle, red – TcdB). Necrotic cell death (Sytox orange – red fluorescence) was quantified at 2-, 24- and 48-hours post-TcdB treatment. A) Both phase-groups have equal activation of necrotic cell death at 48-hours post-TcdB treatment in HIOs. B) The D24 phase group has greater necrotic cell death compared to the D36 phase group at 48-hours post-TcdB treatment in kcHIEs. C) The D24 phase group has greater necrotic cell death compared to the D36 phase group at 48-hours post-TcdB treatment in bHIEs. Data represented as Mean ± S.D. of N=4 biological replicates. Enteroid biological replicates were the average value of N=3 technical replicates per biological replicate. All data were normalized to time/phase matched vehicle controls set to 1. *p<0.05, **p<0.01, ***p<0.001. Scale bar = 250µM. All HIOs were generated from iPSC line 72.3. kcHIE data were generated from lines #3, #4, #5 and eHIE. bHIE data were generated from lines #103, #104, #109 and #179. necrotic cell death response to TcdB. (Figure 3.7(A)). Both kcHIEs and bHIEs possessed a circadian phase-dependent necrotic cell death response, having higher necrosis at the D24 phase compared to the D36 phase (Figure 3.7(B/C)). Similar to mouse, significant necrotic cell death in

TcdB treated samples only occurred at 48-hours post-exposure.

Figure 3.8 – Mouse enteroids possess transcriptomic rhythms. Mouse enteroids, isolated and combined from two separate C57BL/6J mice, were collected every two-hours over two-circadian cycles. Transcript expression, reported as transcripts per million (TPM), at each timepoint was quantified by RNA-sequencing. Rhythmic transcript expression was detected using MetaCycle 2D. A) Heatmap displaying the phase distribution of rhythmic transcripts in mouse enteroids. B) Bmal1 (black) and Rev-erbα (grey) have rhythmic expression profiles with proper phase-distributions supporting the presence of a functional TTFL in the mouse RNA-seq dataset. C) qRT-PCR (green) was used to validate RNA-seq rhythmic expression profiles (blue) of Bmal1 and Rev-erbα. HIOs and tamoxifen treated Bmal1Fx/Fx-EsrCRE enteroids both lack clock activity and a circadian phase-dependent response to TcdB. All mouse and human enteroids that do possess clock activity respond to TcdB with phase-dependent necrotic cell death that is present only at

48-hours post-TcdB treatment. These data provide a cross-species validation of the indispensable role of the intestinal circadian clock in promoting a phase-dependent response to TcdB.

Interestingly, the TcdB responsive phase, phase with greater TcdB induced necrosis, was opposite in mouse and human enteroids. Both PER2::LUC and non-tamoxifen treated Bmal1Fx/Fx-

EsrCRE mouse enteroids had greater necrosis in the D36 phase group whereas bHIEs and kcHIEs had greater necrosis in the D24. Mice are nocturnal and human’s diurnal. Our finding of

Figure 3.9 – Human biopsy derived enteroids possess transcriptomic rhythms. Human biopsy derived enteroids from patient biopsy #103 were collected every two- hours over two-circadian cycles. Transcript expression, reported as transcripts per million (TPM), at each timepoint was quantified by RNA-sequencing. Rhythmic transcript expression was detected using MetaCycle 2D. A) Heatmap displaying the phase distribution of rhythmic transcripts in human enteroids. B) Bmal1 (black) and Rev-erbα (grey) have rhythmic expression profiles with proper phase-distributions, supporting the presence of a functional TTFL in the bHIE RNA-seq dataset. C) qRT-PCR (green) was used to validate RNA-seq rhythmic expression profiles (blue) of Bmal1 and Rev-erbα. anti-phasic TcdB responsiveness could indicate that behavioral activity patterns may be hardwired within peripheral circadian clocks and maintained within in vitro intestinal enteroids. We performed timecourse RNA-sequencing of mouse and human derived enteroids to provide a mechanistic link to the circadian phase-dependent response to TcdB and to explore for species- dependent variability in circadian gene expression in enteroids.

3.4.2 Mouse and human enteroids possess transcriptome wide rhythmic gene

expression and out-of-phase expression of core clock genes and Rac1

Figure 3.10 – Rhythmic transcript phase, in enteroids, is species-dependent. Transcript expression profiles were taken from mouse (red) and human (black) enteroid RNA-seq datasets. A) Bmal1 and Reverbα expression is phase-delayed by ≈6- to 8-hours in human enteroids compared to mouse enteroids. B) Rac1 expression is anti-phasic between human and mouse enteroids. The mouse and human enteroid circadian transcriptome have yet to be published. We therefore collected mouse and human enteroids in timecourse format, every 2-hours over two complete circadian cycles. Transcriptomic expression values were attained by RNA-sequencing.

The rhythmic gene detection algorithm MetaCycle 2D186 was used to generate a comprehensive list of all oscillating genes within both timecourses. Using this approach, we found that up to 20% and 8% of genes are rhythmically expressed in mouse and human enteroids, respectively.

Interestingly, nearly every rhythmic transcript is phase-aligned in the mouse enteroid timecourse with most transcripts peaking at 40- and 64-hours post-synchronization (Figure 3.8). The large

number of in-phase oscillating pathways is atypical compared to other circadian transcriptome analyses80,189,205,206 and may indicate the transcriptomic oscillations are driven by a non-circadian mechanism. Rev-erbα expression is however 8-hours phase-delayed compared to Bmal1, which was confirmed by qRT-PCR analysis of the RNA-seq samples, showing the core clock is likely functional in our mouse enteroid timecourse (Figure 3.8). The phase distribution of rhythmic transcripts in the human timecourse is diverse and similar to previously reported circadian transcriptome heatmaps (Figure 3.9). Like the mouse timecourse we found both Bmal1 and Rev- erbα significantly oscillating with proper phase distributions which was supported by qRT-PCR follow-up, indicating an active clock (Figure 3.9). Given the TcdB responsive phase was opposite in mouse and human enteroids we sought to use these RNA-seq data to identify a species- dependent out-of-phase rhythmic TcdB target that is likely driving the necrotic cell death phenotype in response to TcdB exposure.

Proof-of-concept for out-of-phase rhythmic activity between mouse and human enteroids are shown with out-of-phase oscillations of Bmal1 and Rev-erbα. Human expression of either gene is phase-delayed by approximately 6- to 8-hours compared to mouse (Figure 3.10 (A)). The necrotic TcdB response could be driven by rhythms in small GTPases, TcdB receptors or NADPH complex elements. We searched our mouse and human datasets for rhythmic expression of all of these elements and found rhythmic expression of TcdB targets was species-dependent. All three Rho GTPases (Rac1, RhoA and Cdc42), multiple TcdB receptors (Fzd1, Fzd2, Fzd7 and

Pvrl3) and none of the NADPH complex elements were rhythmic in mouse enteroids. In human enteroids the only rhythmic GTPase was Rac1, and the only rhythmic TcdB receptor was Cspg4.

The NADPH oxidase complex elements Cyba and Noxo1 were also found to be rhythmic in human enteroids, unlike mouse. Rac1 was the only shared rhythmically expressed TcdB target in both timecourses. Interestingly, unlike Bmal1 and Rev-erbα which are ≈8-hours phase-delayed from mouse to human, Rac1 expression appears almost perfectly anti-phasic between mouse and human enteroids (Figure 3.10 (B)). The strong anti-phasic profiles of Rac1 expression between

mouse and human enteroids, coupled with the lack of other shared rhythmically expressed TcdB targets strongly suggests Rac1 is mediating the phase-dependent response to TcdB.

3.5 Conclusions and discussion

In this chapter, we have demonstrated that both mouse and human enteroids respond to

Clostridium difficile toxin B in a circadian phase-dependent fashion. Mouse enteroids isolated from PER2::LUC and Bmal1Fx/Fx-EsrCRE, without tamoxifen treatment, mice had greater necrotic cell death at the D36 phase compared to the D24 phase. Tamoxifen treated arrhythmic Bmal1Fx/Fx-

EsrCRE enteroids lost the circadian phase-dependent response, having similar necrotic cell death at both phases tested. Analogous findings have been made in bone marrow-derived macrophages (BMDMs) following exposure of the protozoan Leishmania major. BMDMs had greater L. major attachment at 32-hours compared to 20-hours post-serum synchronization.

Bmal1KO BMDMS lacked a circadian phase-dependent attachment of L. major119.

TcdB mediated uptake of necrotic and apoptotic cell death dyes was significantly greater than vehicle treated controls after 48-hours of exposure. Necrotic cell death was however the only cell death mechanism initiated in a circadian phase-dependent manner. A recent study using mouse enteroids concluded that apoptosis was the primary cell death response to TcdB, occurring at 16-hours post-exposure207. In that study, however, n-acetyl cysteine (NAC) was added to the mouse enteroid culture media. NAC has been reported to protect HeLa and Caco-2 cells from reactive oxygen species mediated necrotic cell death following TcdB addition153. In our study we found 10ng/mL (37pM) of TcdB was sufficient to induce a TcdB necrotic or apoptotic response.

Saavedra and colleagues utilized 1,000ng/mL (3,700pM) of TcdB207. The presence of NAC in their culture media coupled with the 100X difference in TcdB concentration are a clear diverging point between our study and theirs. Both reports do show that apoptotic cell death is initiated following

TcdB administration with their study clearly indicating inhibition of the apoptotic response can

serve a protective role against TcdB207. However, we found, using two separate mouse enteroid lines, necrotic cell death initiation is activated following TcdB exposure and required for a circadian phase-dependent response to TcdB. We extended upon the mouse enteroid findings by testing multiple 3D human intestinal samples for a similar phenotype. kcHIEs and bHIEs both responded to TcdB in a circadian phase-dependent fashion whereas arrhythmic HIOs had a phase-independent response. These data offer strong support for a novel circadian clock mediated circadian phase-dependent necrotic cell death response to TcdB. Cross species analysis served to validate independent findings made in separate 3D intestinal sample types.

Arrhythmic mouse (tamoxifen Bmal1Fx/Fx-EsrCRE) and human (HIOs) intestinal samples lacked a phase-dependent response to TcdB, acting as negative controls that support the indispensable role of the clock in driving the phase-dependent response to TcdB. Enteroids from two separate mouse lines (PER2::LUC, non-tamoxifen Bmal1Fx/Fx-EsrCRE) responded to TcdB with greater necrotic cell death at the D36 phase. Similarly, two separate human enteroid samples (kcHIEs and bHIEs) responded with a phase-dependent necrotic response, but at the D24 phase.

Enteroids thus appear to maintain the phase of in vivo behavioral patterns (mice nocturnal, human’s diurnal) from the source they were isolated from. We utilized our RNA-sequencing analysis to investigate this claim.

Both mouse and human enteroids present robust circadian transcriptomes. Direct comparison of the RNA-sequencing data found mouse and human enteroids have out-of-phase oscillating transcripts. The core clock genes Bmal1 and Rev-erbα were phase-delayed in human enteroids compared to mouse and Rac1 expression was anti-phasic. When all three Rho small

GTPases (RAC1, RHOA and CDC42) are individually knocked down in Caco-2 cells, only

RAC1KD was protective against TcdB damage153. Although glucosylation of all three GTPases has been implicated in cytotoxic TcdB effects, this finding indicates RAC1 may be serving a primary role. We therefore hypothesize that RAC1 has a central role in driving the observed species-specific phase-dependent response to TcdB. Rac1 attains peak expression at ≈36-hours

post-synchronization in mouse enteroids whereas in humans Rac1 reaches peak expression at

≈28-hours post-synchronization. The peak expression of Rac1 therefore correlates with TcdB responsiveness being greatest at 36- and 24-hours post-synchronization in mouse and human enteroids, respectively. TcdB inactivates Rac1 via glucosylation to induce cellular cytoskeleton breakdown.

To our knowledge, these are the first data that correlate out-of-phase in vivo behavioral patterns with an out-of-phase in vitro pathogenic response. Humans present multiple chronotypes

(morning larks and night owls) along with inter-individual differences in genetic makeup that cannot be recapitulated using inbred mouse models housed under constant environmental conditions. Future effort should be dedicated to understanding how inter-individual differences in clock phase may cause subtle differences in the phase of the pathogenic response in vitro.

Further, aging has been implicated to alter core clock function208 and regulation of downstream clock regulated genes209. Either of these topics can be readily explored in vitro using the bHIE system. How age-dependent changes to the clock may influence the circadian phase-dependent phenotypes described here is discussed further in section 4.2.

Enteroid circadian clock function could be either promoting a TcdB responsive or TcdB protective phase. Direct support for a clock mediated protective phase is offered by the Bmal1Fx/Fx-

EsrCRE enteroids. Non-tamoxifen treated Bmal1Fx/Fx-EsrCRE have a significant increase, compared to vehicle control, in necrotic cell death in only the D36 phase-group at 48-hours post-

TcdB. Tamoxifen treated Bmal1Fx/Fx-EsrCRE have a phase-independent increase in necrotic cell death over vehicle control in both the D36 and D24 phase groups. These data indicate a functional clock may prevent TcdB induced necrosis in the D24 group of non-tamoxifen treated, rhythmic,

Bmal1Fx/Fx-EsrCRE mouse enteroids. bHIEs also have significant necrosis, versus vehicle control, in only one phase group (D24) at 48-hours post exposure. PER2::LUC mouse enteroids show an earlier TcdB response at 24-hours post-TcdB in only their D36 phase-group. Mechanistic understanding of what is driving the TcdB phase-response is required to investigate the protective

vs. responsive hypothesis further. We have provided a correlative association of rhythmic Rac1 in mediating the phase-response to TcdB. Evaluating the phase-response to TcdB following functional manipulations (overexpression and knockdown) of rhythmically expressed TcdB targets, such as Rac1, will provide insight on how the clock primes the enteroid response to TcdB.

Insights such as these could provide translatable implications for chronotherapeutic dosing of C. diff therapies.

4. General discussion and future directions

Ontogeny of the intestinal circadian clock was determined using differentiation induced

HIOs, kcHIEs and bHIEs. Differentiation of PSCs into HIOs was not sufficient to initiate robust circadian rhythms. kcHIEs derived from HIOs matured in NSG mice had circadian clock activity similar to bHIEs. Many hypotheses could be proposed to explain these phenotypes and we have focused on two: a maturity and a circadian hypothesis. The maturity hypothesis assumes that

HIOs require maturation of their epithelium to establish robust rhythmicity following in vitro differentiation. This hypothesis aligns with the lack of circadian rhythms in rat colonic tissue at a similar embryonic age as HIOs177. Rhythmic transplanted HIOs also have enhanced expression of multiple markers of intestinal maturity109,110 compared to arrhythmic HIOs. The circadian hypothesis assumes that while embedded under the mouse kidney capsule, transplanted HIOs are exposed to rhythmic entrainment signals that drive their TTFL to become active. Support for the circadian hypothesis comes from experiments detailing how maternal circadian disruption prompts arrhythmicity in offspring during the perinatal period164. Future directions to test these hypotheses are outlined in section 4.1.

Intestinal circadian clock involvement in the host pathogenic response was shown by challenging multiple organoid and enteroid samples with TcdB from Clostridium difficile. 3D intestinal samples that possessed clock activity responded to TcdB in a circadian phase-

dependent manner. Aging attenuates circadian behavioral rhythms210 and molecular clock activity211. It is intriguing to postulate whether age-related changes to the intestinal clock could mitigate the circadian phase-dependent response to TcdB. Aging is an established risk factor for

Clostridium difficile infection (CDI)212. Whether age-dependent decline in clock function predisposes elderly individuals to CDI remains unknown. The presence of robust clock activity within bHIEs make it an enticing model for evaluating the connection between aging, circadian rhythms, and CDI. Future directions for answering these questions are outlined in section 4.2.

4.1 Identification of circulating circadian signals that establish circadian clock

function in the developing intestine

4.1.1 In vitro activation of the HIO circadian clock

In this study, HIOs were used immediately at the end of the 35-day differentiation protocol.

Long term maintenance of HIOs, up to 140-days, in vitro is possible103. Extended in vitro culture offers the opportunity to manipulate HIOs following complete differentiation. During this extended time, they can be tested for a response that mimics either SCN or food entrainment. Recent evidence indicates that insulin and insulin like growth factor-1 (IGF-1) mediate clock entrainment from food intake. Food intake prompts a hyperglycemic state that results in insulin release and thus activity of IGF-1 which works through mammalian target of rapamycin complex 1 (mTORC1) to amplify PER2 which resets peripheral clocks by inhibition of BMAL1:CLOCK transcriptional activation160. SCN synchronization of peripheral clocks can occur through neuronal and endocrine mechanisms73,74,76. Dexamethasone (DEX) is a potent synchronization stimulus for in vitro cells that acts as a mimic of SCN derived glucocorticoid synchronization204. Host mouse vascularization of transplanted HIOs establishes a permissive environment for HIO synchronization via glucocorticoid signaling from the SCN or insulin following food consumption. Periodic administration of insulin or DEX to HIOs can be used to test if the HIO clock can be activated by

Figure 4.1 – Environmental manipulations to evaluate host circadian input on clock development in transplanted HIOs. Designs for testing whether host mice entrain transplanted HIO samples. Top bar is light-dark (white = light, black = dark) cycle and bottom bar is timing of food availability (continuous bar = ad lib, broken bars = restricted food availability). A) Traditional mouse housing conditions, 12:12 LD cycle with ad-lib food availability. kcHIEs generated for previous tests were maintained in mice housed under this paradigm. B) Mice housed under constant darkness will enter a circadian free-running state to promote endogenous SCN activity. Top group will be offered food ad-lib, bottom group will have food restricted (RF) to 8-hours within the dark phase of the previous LD schedule only. C) Mice housed under constant lighting to disrupt SCN circadian entrainment. Top group will be offered food ad-lib, bottom group will have food restricted (RF) to 8-hours within the dark phase of the previous LD schedule only. in vitro circadian entrainment. HIOs would be dosed with either DEX or insulin every day, at the same time, beginning from day-35 of the HIO differentiation protocol. Similar in vitro designs successfully synchronized SCN neurons with gamma-aminobutyric acid (GABA)63 and amplified the clock of B16 melanoma derived cells with DEX213. HIOs treated with a single dose of DEX produced low amplitude and unsustainable oscillations of Bmal1-luc. Repeated administration of either DEX or insulin could, however, initiate the intestinal circadian clock by mimicking rhythmic host entrainment cues. If daily timed administration of DEX or insulin is capable of promoting robust clock activity in vitro, we could correlate that to the number of days it takes to activate the clock during in vivo maturation of HIOs in NSG mice. In vivo kidney capsule transplantation of

HIOs would allow us to validate the in vitro entrainment findings on the length of time, and signaling mechanisms that activate early intestinal clock activity.

4.1.2 Evaluation of host circadian input on clock activation in transplanted HIOs

Transplantation and maturation of HIOs in NSG mice was necessary to gain kcHIEs that possessed robust clock rhythms. The mechanism that established the intestinal circadian clock in NSG mice remain elusive. All of the offspring born from mothers with SCN lesion164, exposed to constant light172 or transgenic arrhythmicity168 successfully establish circadian behavioral activity rhythms. However, pups from the circadian disrupted mothers showed delayed onset of activity rhythms compared to control164. These data indicate maternal derived circadian signals serve a role in coordinating perinatal, early, circadian clock activity in offspring. HIO transplantation can extend upon these findings by evaluating the role of in vivo host mouse signaling on human clock ontogeny in the transplanted HIO. Although not connected by a placenta, transplanted HIOs become vascularized by the host mouse110 and are thus exposed to the host mouse circulation. Blood born circadian signaling factors are generated from the SCN74 or food intake160. Our current data show that when transplanted into mice, HIOs give rise to rhythmic kcHIEs. These mice were housed under typical 12-hour light dark (LD) cycles and provided food ad libitum (AL). By manipulating the host’s environmental lighting and food availability, one can investigate which host circadian signals influence clock activation during HIO transplantation. Mice housed under constant darkness (DD) manifest a free-running state without entrainment from external time cues214. Constant light exposure (LL) disrupts synchronized neuronal activity within the SCN171 inhibiting SCN synchronization of peripheral clocks215. In rats,

30-days of LL attenuated rhythmic gene expression of most core clock genes within the liver and duodenum. Food restriction to 6-hours per circadian cycle in LL conditions, however, rescued rhythmicity of core clock genes215. In other words, nutrient intake is a potent zeitgeber for peripheral clocks that can also be tested for its input to clock activation in transplanted HIOs.

Manipulating the environment of HIO transplanted mice would provide novel insights regarding

how environmental signals influence the host’s ability to generate intestinal samples that possess clock activity.

Mice would be housed under two different lighting conditions, DD and LL, and two different food paradigms, ad libitum (AL) and restricted food (RF). This experimental set up will provide four independent conditions for mice: DD-AL, DD-RF, LL-AL, and LL-RF (Figure 4.1). Food restricted mice in DD or LL would have food availability limited to 8-hours during the dark phase of the previous 12:12 LD cycle. The in vitro experiments in 4.1.1 will inform on the number of days it takes for HIO clock entrainment (i.e. robust HIO clock activity is observed after 10-days of repeated dosing of HIOs with insulin). We will transplant HIOs into NSG mice for an equivalent number of days under the environmental conditions outlined in Figure 4.1. Human tissue will then be isolated to generate kcHIEs. kcHIEs will be tested for clock activity by transduction with the

Bmal-luc reporter. The presence of clock activity in kcHIEs isolated from mice housed under DD-

AL and DD-RF would indicate SCN entrainment whereas kcHIEs derived from DD-RF and LL-RF groups would indicate food entrainment. Transplanted HIOs in DD-AL mice will be solely entrained by the SCN and LL-RF housed mice will only receive entrainment cues from rhythmic food intake. Comparison of kcHIEs generated from mice housed under DD-AL and LL-RF conditions will provide the clearest evidence of whether the SCN, rhythmic food intake or both can entrain clock activity in transplanted HIOs. LL-AL mice will serve as a negative control.

Constant light abolishes SCN synchronization of peripheral tissues215. During ad lib feeding mice primarily consume food during the active (dark) phase but also have food intake, to a lesser extent, during the light phase. The lack of scheduled food consumption during ad lib feeding mitigates the possibility for food entrainment, particularly under LL conditions when rhythmic behavioral activity is disrupted. Lack of clock activity in kcHIEs isolated from LL-AL mice would support entrainment as a requirement for circadian rhythm activation in transplanted HIOs.

6-weeks of maturation under the kidney capsule of NSG mice successfully matures

HIOs110. After 6-weeks, transplanted HIOs contain well-defined crypt domains that can be isolated

to establish kcHIEs following the same protocol used for establishing human biopsy enteroids101.

Generation of kcHIEs prior to 6-weeks may prove difficult if crypt domains are not mature enough to be isolated and embedded into a 3D Matrigel dome. This limitation would require in vivo testing for the presence of a functional clock at earlier timepoints post-kidney capsule transplantation. In this case, mice would be similarly housed under the environmental conditions in Figure 4.1 and sacrificed at the post-transplantation timepoint when HIO clock activity is hypothesized to be active. Sacrifice would occur at two timepoints 12-hours apart. Immunofluorescence staining of the human tissue would be used to detect differential clock gene presence at the two out-of-phase timepoints. Combining in vitro and in vivo entrainment findings would provide strong support for the indispensable role of entrainment in generating robust clock function in immature tissues.

In vitro generated HIOs and in vivo matured kcHIEs offer an ideal system for understanding the basic mechanisms of intestinal development. These two processes can be tightly regulated and manipulated to intricately test developmental hypotheses. Further work on the HIO/kcHIE systems will be necessary to uncover the molecular mechanisms regulating the maturation of HIOs in NSG mice. Knowledge of the processes driving the in vivo maturation of

HIOs will facilitate the development of methods to mature HIOs in vitro, circumventing the need for in vivo transplantation in NSG mice. We have proposed future experiments designed to test in vitro maturation of circadian rhythms using insulin or dexamethasone. In vitro findings will be validated by altering host mouse entrainment in vivo via environmental manipulations. These experiments operate under the assumption that synchronization, rather than cellular maturation, can serve as the dominant initiating factor for intestinal circadian rhythms. If in vivo maturation, rather than entrainment, of transplanted HIO samples drives the development of intestinal clock activity, a next step would be to identify which cell types, when matured, activate the HIO clock.

Our lab has previously used mouse enteroids to show rhythmic WNT release from Paneth cells coordinate rhythmic cell cycle progression in intestinal stem cells and progenitor cells92. Paneth cells are also indispensable for the generation of mouse enteroids100. These insights point to

Figure 4.2 – Network interaction between the core circadian clock loop and SIRT1/NAD+. Points of connection between the circadian clock TTFL (green) and a simplified SIRT1/NAD+ loop (blue). Black arrows indicate activating interactions, red dashed lines indicate repressive interactions.

Paneth cells as an ideal starting point for studies investigating the influence of individual cell maturation state on early clock function in HIOs. In vivo SOX9 regulates the number and differentiation status of Paneth cells216. In vitro WNT/R-spondin concentrations serve similar functions within mouse intestinal enteroids217. Modulation of either signaling pathway may be useful in promoting Paneth cell maturation at an early timepoint within HIOs to induce circadian rhythms.

4.2 Characterization of intestinal clock aging and its influence on the response

to Clostridium difficile toxin B.

Figure 4.3 – Preliminary insights on the influence of donor age on the bHIE circadian phase-dependent response to TcdB. D36 and D24 phase-groups were tested for a circadian phase-dependent necrotic cell death response following TcdB, 5,000ng/mL (18.5nM), addition. bHIE donors were separated into different age groups: Young (<30 years, N=3: #103, #111, #179), Middle-aged (30 years < X < 60 years, N=1: #109) and Old (60< years, N=2: #104, #110). A) Young patients have greater TcdB induced necrotic cell death in the D24 phase group compared to the D36 phase group, at 48-hours post-TcdB addition. B) A middle-aged donor had greater necrotic cell death in the D24 group at both 24- and 48-hours post-TcdB addition. C) There was no phase-dependent necrotic cell death response following TcdB exposure at any timepoint in bHIEs from old donors. Mouse enteroids derived from 3-6 month-old mice, bHIEs derived from patient donors of multiple ages and kcHIEs all demonstrate a circadian phase-dependent necrotic cell death response to C. diff toxin B. Older individuals are known to be more susceptible to C. diff infection

(CDI)212 but fundamental mechanisms driving this phenotype have yet to be systematically evaluated. Prior to bHIEs there was a lack of in vitro aging models for gaining insight on C. diff’s differential interaction with the intestine of young and aged humans. In vivo, aged mice have been

used to model the age-dependent response to C. diff. Exposure of aged mice to C. diff resulted in more extensive intestinal damage and a longer time to recovery compared to young mice. Age- dependent attenuation of the immune response and an altered microbiota profile were the causal links to the greater pathology218,219. In vivo work can evaluate how multiple systems (i.e. immune/microbiota) interact during the pathogenic response. However, it is difficult to uncover the detailed molecular mechanism that cause in vivo phenotypes. bHIEs can be isolated from donors spanning multiple age ranges and better recapitulate in vivo intestinal biology compared to other in vitro system. These characteristics make bHIEs an attractive in vitro platform for studying age-dependent responses to C. diff and its toxins. Aging also prompts changes to the circadian clock and its downstream targets208,209. Whether age-dependent circadian clock phenotypes are recapitulated in bHIEs, and how an aging clock influences the circadian phase- dependent response to C. dfif toxin B remains unknown.

The proposed future direction is to use bHIEs to determine how aging alters the intestinal clock, and its downstream targets. Similar to the transition from Chapter 2 to Chapter 3 here, we will extend upon this characterization by testing how age-dependent changes to the intestinal clock alter the pathogenic response to TcdB. We hypothesize that aging alters the expression of

CCGs, including TcdB targets, and attenuates the circadian phase-dependent response to TcdB.

Aging is known to disrupt the circadian system which behaviorally manifests with altered sleep/wake and activity cycles210. How aging impacts the molecular transcriptional translational feedback loop (TTFL) and clock regulated targets is currently an active area of investigation.

Attenuation of core clock gene function in central208 and peripheral211 clocks was initially proposed to explain age-related decline in circadian function. A more recent comparison of young and aged mouse liver has challenged these findings, showing that it is not core clock genes differentially expressed between young and aged livers but instead it is what the core clock genes are regulating that changes209. Whether aging alters the core clock or downstream targets, age- dependent loss of SIRT1 activity has been the proposed mechanistic target to explain the age-

related clock phenotypes208,209. SIRT1 and NAD+ (SIRT1’s co-factor) signaling interact with the circadian TTFL. BMAL1:CLOCK rhythmically regulate the transcription of nicotinamide phosphoribosyltransferase (NAMPT)220,221. NAMPT is the rate-limiting enzyme in the NAD+ salvage pathway, rhythmic expression of NAMPT drive oscillations in the level of NAD+. Direct evidence for clock regulation of these rhythms was shown with the loss of rhythmic NAMPT/NAD+ in circadian arrhythmic ClockΔ19 mice221. NAD+ activates SIRT1 in a dose-dependent manner, and rhythmic NAD+ is likely to be contributing to rhythmic SIRT1 activity222. SIRT1 feeds back on the core clock through histone deacetylation of Bmal1 at H3 Lys9/Lys14222 and by promoting PER2 expression223. This bi-directional relationship between SIRT1/NAD+ and the circadian clock generates another feedback loop outside of the canonical TTFL (Figure 4.2). The histone deacetylation profile and number of rhythmically expressed genes is attenuated in aged mouse liver samples. Upon calorie restriction, a known enhancer of Sir2 (SIRT1 homolog) and NAD+224, the histone deacetylation profile and number of rhythmically expressed genes is enhanced in the aged liver209. These findings indicate aging attenuates circadian clock regulation of CCGs but also show that the aging phenotype can be reversed by enhancing SIRT1 and NAD+ activity.

Our preliminary data show that bHIEs isolated from aged donors lack a circadian phase- dependent response to TcdB (Figure 4.3). These data suggest one or more rhythmic TcdB target genes may have altered expression profiles during aging. To test this hypothesis, we would perform timecourse RNA-sequencing on bHIEs isolated from an aged donor. Comparison of RNA- sequencing from aged and young bHIEs would establish whether aging influences the rhythmic transcriptome of the human intestinal clock. Comparison of TcdB target expression in young and aged bHIEs would be used to explain the lack a phase-dependent response to TcdB in aged samples. The next goal would be to test whether circadian clock function in aged bHIEs can be activated, as was shown in mouse liver following calorie restriction209. Rescue of aged intestinal clock function would then be functionally validated by testing for a rescue of the circadian phase- dependent response to C. diff toxins in aged, clock recovered, bHIEs. 77

Primary candidates to rejuvenate circadian rhythms in aged bHIEs include NAD+/SIRT1 activators and the flavonoid, nobiletin (NOB). NAD+ is incapable of being imported into a cell and must have its cellular pools enhanced by manipulation of its salvage pathway225. Nicotinamide mononucleotide (NMN) and nicotinamide (NAM) are both substrates in the NAD+ salvage pathway. NAM is converted to NMN by NAMPT226 and NMN gets converted to NAD+ by

NMNAT2227. Chronic NMN treatment reduced negative age-related phenotypes in mice228 and

NAM treatment enhances SIRT1 function in skeletal muscle of aged rats via NAD+229. Enhancing either NAM or NMN levels could indirectly activate the circadian clock by promotion of SIRT1 activity via increasing NAD+ pools. NAM is also a SIRT1 inhibitor when added at high concentrations230. Titration of NAM in bHIEs could offer translational insights on the optimal concentrations that provide clock benefit without inhibiting SIRT1 within the human intestine.

Resveratrol and SRT1720, direct activators of SIRT1, could also be tested for their influence in rescuing circadian rhythms in bHIEs derived from aged patients. Both compounds enhance either lifespan and/or healthspan through activation of SIRT1 in vivo231,232. Nobiletin, a natural flavonoid, is a circadian clock activator233. NOB treatment in aged mice is also protective against the negative metabolic effects of a high fat diet in skeletal muscle234. Each of these candidates could be easily screened for efficacy in enhancing the function of circadian rhythms in aged bHIEs using bHIEs transduced with the Bmal1-luc clock reporter. Once identified, the compound would be administered periodically to amplify the expression of CCGs. Clock-rescued aged bHIEs would be challenged with TcdB to test for the rescue of a phase-dependent response. Negative controls would be addition of either EX-527 or FK866 to inhibit SIRT1235 or NAD+220,236, respectively.

Similar treatment could be carried out in bHIEs derived from young donors. In young bHIEs, clock amplifiers could promote a more robust phase-response whereas EX-527 or FK866 attenuation of SIRT1 or NAD+ would mitigate the phase-dependent response.

4.3 Summary 78

We have successfully used the HIO/kcHIE system to characterize ontogeny of the intestinal clock in a system that mimics human embryonic/fetal intestinal development. These data are the first to provide insight into clock ontogeny in human intestinal tissue. We found that differentiation of induced pluripotent stem cells (iPSCs) into multicellular HIOs is incapable of producing circadian rhythms. These data contrast a study that showed differentiation of human embryonic stem cells into cardiomyocytes resulted in clock activity117. Instead, HIOs required kidney-capsule transplantation to generate robust circadian rhythms. We are the first to couple in vitro differentiation and in vivo maturation to establish circadian clock activity. Here we limited our investigation to clock characterization of these culture systems. The HIO/kcHIE system offers immense potential to understand the developmental mechanisms that explain the phenotypic characterization we have reported. We have outlined two future directions (section 4.1) aimed at determining if circadian entrainment or tissue maturation activates clock activity during intestinal development. The intestinal circadian clock regulates numerous biological processes within

IECs40,92,195 and has a bidirectional relationship with the microbiome194,195 that has been implicated in regulation of biological functions outside the intestinal system (i.e. liver detoxification)196,237.

Knowledge of when and how the intestinal circadian clock becomes active will provide an understanding of the developmental timepoint when these processes begin their rhythmic activity.

Using our characterization of clock activity in HIOs, kcHIEs and bHIEs we found that the presence of a functional clock directs the intestinal epithelium to respond to Clostridium difificle toxin B in a circadian phase-dependent manner. Mouse enteroids also possessed a circadian phase-dependent response. Interestingly, the mouse TcdB responsive phase was opposite of the human responsive phase. These data indicate that species behavioral activity rhythms may be maintained within in vitro enteroid samples. RNA-sequencing analysis of timecourse collected samples found up to 20% and 8% of genes are rhythmically regulated in mouse and human enteroids, respectively. Comparison of the two datasets showed that core clock genes were out- of-phase. Further, the only TcdB target rhythmically expressed in both samples was Rac1 which

was anti-phasic between mouse and human enteroids. Peak Rac1 expression occurred at 28- hours in bHIEs and at 36-hours in mouse enteroids. This correlated with greater necrotic cell death in the D24 phase in bHIEs and D36 phase in mouse enteroids. From these insights we preliminary conclude that species-dependent variation in the phase of clock gene expression drive out-of-phase expression of Rac1 which is responsible for the out-of-phase response to TcdB.

Aging has been associated with attenuated core clock activity208 and fewer clock regulated genes209. Aged populations are also the most susceptible to CDI212. In section 4.2 we outlined methods to use bHIEs to investigate the role of aging on the clock and subsequently the response to C. diff toxin B. Whether an anti-CDI treatment could be administered in a chronotherapeutic manner has yet to be tested. Investigation into aged clock activity and response to TcdB will inform on the potential validity of this approach. This underscores the promise of bHIEs as a human relevant intestinal model that possesses robust clock activity and can be isolated from multiple aged donors. In an idealized setting the ability to generate bHIEs, or other organ mini- structures, would be easy and painless in a clinical setting. Robust isolation of these samples would provide a simple means for evaluating a patient’s intestinal function and also their clock function/phase. With this knowledge we could understand the phase of an individual’s clock which is subtly different between individuals (i.e. morning larks vs. nigh owls). Knowing an individual’s intestinal clock phase could then inform on when they will be most susceptible to the damaging effects of something similar to C. diff TcdB. Using this approach, a personalized treatment regimen could be recommended to enhance the efficacy of interventions and improve outcomes.

In this thesis we have characterized clock activity across multiple 3D intestinal sample types and used that characterization to establish an unknown input of the circadian system on the response to TcdB. We have preliminarily correlated rhythmic Rac1 expression as the mechanistic link to the circadian phase-dependent response to TcdB. Finally, we have provided preliminary insights that bHIEs derived from young, middle-aged and old donors have different circadian phase-dependent responses to TcdB. I am hopeful these data will serve as proof of principle for

the use of bHIEs as an model to understand inter-individual differences in intestinal circadian clock function and how that influences their personal response to either pathogens or treatment.

With such understandings we can harness bHIEs to their full potential; to gain insights related to offering the right treatment to the right person and, most intriguingly, at the right time.

References:

1. Pittendrigh, C. S. On Temperature Independence in the Clock System Controlling Emergence Time in Drosophila. Proc. Natl. Acad. Sci. 40, 1018–1029 (1954). 2. Konopka, R. J. & Benzer, S. Clock Mutants of Drosophila melanogaster. Proc. Natl. Acad. Sci. 68, 2112–2116 (1971). 3. Reddy, P. et al. Molecular analysis of the period locus in Drosophila melanogaster and identification of a transcript involved in biological rhythms. Cell 38, 701–710 (1984). 4. Bargiello, T. A., Jackson, F. R. & Young, M. W. Restoration of circadian behavioural rhythms by gene transfer in Drosophila. Nature 312, 752–754 (1984). 5. Stal, L. J. & Krumbein, W. E. Nitrogenase activity in the non-heterocystous cyanobacterium Oscillatoria sp. grown under alternating light-dark cycles. Arch. Microbiol. 143, 67–71 (1985). 6. Colon-Lopez, M., Sherman, D. M. & Sherman, L. A. Transcriptional and Translational Regulation of Nitrogenase in Light-Dark- and Continuous-Light-Grown Cultures of the Unicellular Cyanobacterium Cyanothece sp. Strain ATCC 51142. Bacteriology 179, 4319–4327 (1997). 7. Ishiura, M. et al. Expression of a Gene Cluster kaiABC as a Circadian Feedback Process in Cyanobacteria. Science (80-. ). 281, 1519–1523 (1998). 8. Dunlap, J. C. Molecular Bases for Circadian Clocks. Cell 96, 271–290 (1999). 9. Nakahira, Y. et al. Global gene repression by KaiC as a master process of prokaryotic circadian system. Proc. Natl. Acad. Sci. 101, 881–885 (2004). 10. Tomita, J., Nakajima, M., Kondo, T. & Iwasaki, H. No Transcription-Translation Feedback in Circadian Rhythm of KaiC Phosphorylation. Science (80-. ). 307, 251–254 (2005). 11. Nishiwaki, T., Iwasaki, H., Ishiura, M. & Kondo, T. Nucleotide binding and autophosphorylation of the clock protein KaiC as a circadian timing process of cyanobacteria. Proc. Natl. Acad. Sci. 97, 495–499 (2000). 12. Iwasaki, H., Nishiwaki, T., Kitayama, Y., Nakajima, M. & Kondo, T. KaiA-stimulated KaiC phosphorylation in circadian timing loops in cyanobacteria. Proc.Natl.Acad.Sci.USA 99, 15788–15793 (2002). 13. Kitayama, Y., Iwasaki, H., Nishiwaki, T. & Kondo, T. KaiB functions as an attenuator of KaiC phosphorylation in the cyanobacterial circadian clock system. EMBO J. 22, 2127– 2134 (2003). 14. Rust, M. J., Markson, J. S., Lane, W. S., Fisher, D. S. & O’Shea, E. K. Ordered Phosphorylation Governs Oscillation of a Three-Protein Circadian Clock. Science (80-. ). 318, 809–812 (2007). 15. Teng, S.-W., Mukherji, S., Moffitt, J. R., de Buyl, S. & O’Shea, E. K. Robust Circadian Oscillations in Growing Cyanobacteria Require Transcriptional Feedback. Science (80-. ). 340, 737–740 (2013). 16. Takahashi, J. S. Transcriptional architecture of the mammalian circadian clock. Nat. Rev. Genet. 18, 164–179 (2017). 17. Feldman, J. F. & Hoyle, M. N. Isolation of circadian clock mutants of neurospora crassa. Genetics 75, 605–613 (1973). 18. McClung, C. R., Fox, B. A. & Dunlap, J. C. The Neurospora clock gene frequency shares a sequence element with the Drosophila clock gene period. Nature 339, 558–562 (1989). 19. Crosthwaite, S. K., Dunlap, J. C. & Loros, J. J. Neurospora wc-1 and wc-2: Transcription, Photoresponses, and the Origins of Circadian Rhythmicity. Science (80-. ). 276, 763–769 (1997). 20. Ballario, P., Talora, C., Galli, D., Linden, H. & Macino, G. Roles in dimerization and blue light photoresponse of the PAS and LOV domains of Neurospora crassa white collar

proteins. Mol. Microbiol. 29, 719–729 (1998). 21. Froehlich, A. C., Loros, J. J. & Dunlap, J. C. Rhythmic binding of a WHITE COLLAR- containing complex to the frequency promotor is inhibited by FREQUENCY. Proc. Natl. Acad. Sci. 100, 5914–5919 (2003). 22. Belden, W. J., Loros, J. J. & Dunlap, J. C. Execution of the Circadian Negative Feedback Loop in Neurospora Requires the ATP-Dependent Chromatin-Remodeling Enzyme CLOCKSWITCH. Mol. Cell 25, 587–600 (2007). 23. Cheng, P., He, Q., He, Q., Wang, L. & Liu, Y. Regulation of the Neurospora circadian clock by an RNA helicase. Genes Dev. 19, 234–241 (2005). 24. Hong, C. I., Ruoff, P., Loros, J. J. & Dunlap, J. C. Closing the circadian negative feedback loop: FRQ-dependent clearance of WC-1 from the nucleus. Genes Dev. 22, 3196–3204 (2008). 25. Rutila, J. E. et al. Cycle is a second bHLH-PAS clock protein essential for circadian rhythmicity and transcription of Drosophila period and timeless. Cell 93, 805–814 (1998). 26. Allada, R., White, N. E., So, W. V., Hall, J. C. & Rosbash, M. A mutant Drosophila homolog of mammalian Clock disrupts circadian rhythms and transcription of period and timeless. Cell 93, 791–804 (1998). 27. Hao, H., Allen, D. L. & Hardin, P. E. A circadian enhancer mediates PER-dependent mRNA cycling in Drosophila melanogaster. Mol. Cell. Biol. 17, 3687–3693 (1997). 28. Taylor, P. & Hardin, P. E. Rhythmic E-Box Binding by CLK-CYC Controls Daily Cycles in per and tim Transcription and Chromatin Modifications. Mol. Cell. Biol. 28, 4642–4652 (2008). 29. Nawathean, P., Stoleru, D. & Rosbash, M. A Small Conserved Domain of Drosophila PERIOD Is Important for Circadian Phosphorylation, Nuclear Localization, and Transcriptional Repressor Activity. Mol. Cell. Biol. 27, 5002–5013 (2007). 30. Curtin, K. D., Huang, Z. J. & Rosbash, M. Temporally regulated nuclear entry of the Drosophila period protein contributes to the circadian clock. Neuron 14, 365–372 (1995). 31. Vosshall, L. B. et al. Block in Nuclear Localization of period Protein timeless by a Second Clock Mutation ,. Science (80-. ). 263, 1606–1609 (1994). 32. Kim, E. Y., Hardin, P. E., Ko, H. W., Yu, W. & Edery, I. A DOUBLETIME Kinase Binding Domain on the Drosophila PERIOD Protein Is Essential for Its Hyperphosphorylation, Transcriptional Repression, and Circadian Clock Function. Mol. Cell. Biol. 27, 5014–5028 (2007). 33. Lee, C., Bae, K. & Edery, I. PER and TIM Inhibit the DNA Binding Activity of a Drosophila CLOCK-CYC/dBMAL1 Heterodimer without Disrupting Formation of the Heterodimer: a Basis for Circadian Transcription. Mol. Cell. Biol. 19, 5316–5325 (1999). 34. Kim, E. Y. & Edery, I. Balance between DBT/CKie kinase and protein phosphatase activaties regulate phosphorylation and stability of Drosophila CLOCK protein. PNAS 103, 6178–6183 (2006). 35. Blau, J. & Young, M. W. Cycling vrille expression is required for a functional Drosophila clock. Cell 99, 661–671 (1999). 36. Luo, W. et al. Old flies have a robust central oscillator but weaker behavioral rhythms that can be improved by genetic and environmental manipulations. Aging Cell 11, 428–438 (2012). 37. Cyran, S. A. et al. vrille, Pdp1, and dClock form a second feedback loop in the Drosophila circadian clock. Cell 112, 329–341 (2003). 38. Dubruille, R. & Emery, P. A plastic clock: How circadian rhythms respond to environmental cues in drosophila. Mol. Neurobiol. 38, 129–145 (2008). 39. Plautz, J. D., Kaneko, M., Hall, J. C. & Kay, S. A. Independent photoreceptive circadian clocks throughout Drosophila. Science (80-. ). 278, 1632–1635 (1997). 40. Karpowicz, P., Zhang, Y., Hogenesch, J., Emery, P. & Perrimon, N. The Circadian Clock

Gates the Intestinal Stem Cell Regenerative State. Cell Rep. 3, 996–1004 (2013). 41. Giebultowicz, J. M. & Hege, D. M. Circadian clock in Malpighian tubules. Nature 386, 664 (1997). 42. Ito, C. & Tomioka, K. Heterogeneity of the peripheral circadian systems in Drosophila melanogaster: A review. Front. Physiol. 7, 1–7 (2016). 43. Schmalen, I. et al. Interaction of circadian clock proteins CRY1 and PER2 is modulated by zinc binding and disulfide bond formation. Cell 157, 1203–1215 (2014). 44. Lee, C., Etchegaray, J.-P., Cagampang, F. R., Loudon, A. S. & Reppert, S. M. Posttranslational Mechanisms Regulate the Mammalian Circadian Clock. Cell 107, 855– 867 (2001). 45. Reischl, S. et al. β-TrCP1-mediated degradation of PERIOD2 is essential for circadian dynamics. J. Biol. Rhythms 22, 375–386 (2007). 46. Siepka, S. M. et al. Circadian Mutant Overtime Reveals F-box Protein FBXL3 Regulation of Cryptochrome and Period Gene Expression. Cell 129, 1011–1023 (2007). 47. Preitner, N. et al. The orphan nuclear receptor REV-ERB controls circadian transcription within the positive limb of the mammalian circadian oscillator. Cell 110, 251–260 (2002). 48. Sato, T. K. et al. A functional genomics strategy reveals rora as a component of the mammalian circadian clock. Neuron 43, 527–537 (2004). 49. Guillaumond, F., Dardente, H., Giguère, V. & Cermakian, N. Differential control of Bmal1 circadian transcription by REV-ERB and ROR nuclear receptors. J. Biol. Rhythms 20, 391–403 (2005). 50. Ruben, M. D. et al. A database of tissue-specific rhythmically expressed human genes has potential applications in circadian medicine. Sci. Transl. Med. 10, 1–8 (2018). 51. Wang, F. et al. Meta-analysis on night shift work and risk of metabolic syndrome. Obes. Rev. 15, 709–720 (2014). 52. Vetter, C. et al. Association between rotating night shiftwork and risk of coronary heart disease among women. JAMA 315, 1726–1734 (2016). 53. Korsiak, J., Tranmer, J., Day, A. & Aronson, K. J. Sleep duration as a mediator between an alternating day and night shift work schedule and metabolic syndrome among female hospital employees. Occup. Environ. Med. 75, 132–138 (2018). 54. Scheer, F. A. J. L., Hilton, M. F., Mantzoros, C. S. & Shea, S. A. Adverse metabolic and cardiovascular consequences of circadian misalignment. Proc. Natl. Acad. Sci. U. S. A. 106, 4453–4458 (2009). 55. Leproult, R., Holmbäck, U. & Van Cauter, E. Circadian misalignment augments markers of insulin resistance and inflammation, independently of sleep loss. Diabetes 63, 1860– 1869 (2014). 56. Panda, S. et al. Melanopsin (Opn4) requirement for normal light-induced circadian phase shifting. Science (80-. ). 298, 2213–2216 (2002). 57. Damiola, F. et al. Restricted feeding uncouples circadian oscillators in peripheral tissues from the central pacemaker in the suprachiasmatic nucleus. Genes Dev. 14, 2950–2961 (2000). 58. Salgado-Delgado, R. C. et al. Shift Work or Food Intake during the Rest Phase Promotes Metabolic Disruption and Desynchrony of Liver Genes in Male Rats. PLoS One 8, 1–11 (2013). 59. Ruby, N. F. et al. Role of melanopsin in circadian responses to light. Science (80-. ). 298, 2211–2213 (2002). 60. Berson, D. M., Dunn, F. A. & Takao, M. Phototransduction by retinal ganglion cells that set the circadian clock. Science (80-. ). 295, 1070–1073 (2002). 61. Moore, R. Y. & Lenn, N. J. A retinohypothalamic projection in the rat. J. Comp. Neurol. 146, 1–14 (1972). 62. Yamaguchi, S. et al. Synchronization of Cellular Clocks in the Suprachiasmatic Nucleus.

Science (80-. ). 302, 1408–1412 (2003). 63. Liu, C. & Reppert, S. M. GABA synchronizes clock cells within the suprachiasmatic circadian clock. Neuron 25, 123–128 (2000). 64. Webb, A. B., Angelo, N., Huettner, J. E. & Herzog, E. D. Intrinsic, nondeterministic circadian rhythm generation in identified mammalian neurons. Proc. Natl. Acad. Sci. U. S. A. 106, 16493–16498 (2009). 65. Welsh, D. K., Logothetis, D. E., Meister, M. & Reppert, S. M. Individual neurons dissociated from rat suprachiasmatic nucleus express independently phased circadian firing rhythms. Neuron 14, 697–706 (1995). 66. Aton, S. J., Colwell, C. S., Harmar, A. J., Waschek, J. & Herzog, E. D. Vasoactive intestinal polypeptide mediates circadian rhythmicity and synchrony in mammalian clock neurons. Nat. Neurosci. 8, 476–483 (2005). 67. Herzog, E. D., Hermanstyne, T., Smyllie, N. J. & Hastings, M. H. Regulating the suprachiasmatic nucleus (SCN) circadian clockwork: Interplay between cell- autonomous and circuit-level mechanisms. Cold Spring Harb. Perspect. Biol. 9, 1–26 (2017). 68. Barca-Mayo, O. et al. Astrocyte deletion of Bmal1 alters daily locomotor activity and cognitive functions via GABA signalling. Nat. Commun. 8, 1–14 (2017). 69. Tso, C. F. et al. Astrocytes Regulate Daily Rhythms in the Suprachiasmatic Nucleus and Behavior. Curr. Biol. 27, 1055–1061 (2017). 70. Brancaccio, M., Patton, A. P., Chesham, J. E., Maywood, E. S. & Hastings, M. H. Astrocytes Control Circadian Timekeeping in the Suprachiasmatic Nucleus via Glutamatergic Signaling. Neuron 93, 1420–1435 (2017). 71. Jones, J. R., Simon, T., Lones, L. & Herzog, E. D. SCN VIP neurons are essential for normal light-mediated resetting of the circadian system. J. Neurosci. 38, 7986–7995 (2018). 72. Pando, M. P., Morse, D., Cermakian, N. & Sassone-Corsi, P. Phenotypic rescue of a peripheral clock genetic defect via SCN hierarchical dominance. Cell 110, 107–117 (2002). 73. Silver, R., LeSauter, J., Tresco, P. A. & Lehman, M. N. A diffusible coupling signal from the transplanted suprachiasmatic nucleus controlling circadian locomotor rhythms. Nature 382, 810–813 (1996). 74. Guo, H., Brewer, J. M., Champhekar, A., Harris, R. B. S. & Bittman, E. L. Differential control of peripheral circadian rhythms by suprachiasmatic-dependent neural signals. Proc. Natl. Acad. Sci. U. S. A. 102, 3111–3116 (2005). 75. Kalsbeek, A., La Fleur, S., Van Heijningen, C. & Buijs, R. M. Suprachiasmatic GABAergic inputs to the paraventricular nucleus control plasma glucose concentrations in the rat via sympathetic innervation of the liver. J. Neurosci. 24, 7604–7613 (2004). 76. Mohawk, J. A., Green, C. B. & Takahashi, J. S. Central and Peripheral Circadian Clocks in Mammals. Annu. Rev. Neurosci. 35, 445–462 (2012). 77. Lamia, K. A., Storch, K. F. & Weitz, C. J. Physiological significance of a peripheral tissue circadian clock. Proc. Natl. Acad. Sci. U. S. A. 105, 15172–15177 (2008). 78. Stokkan, K. a, Yamazaki, S., Tei, H., Sakaki, Y. & Menaker, M. Entrainment of the circadian clock in the liver by feeding. Science 291, 490–493 (2001). 79. Mukherji, A. et al. Shifting eating to the circadian rest phase misaligns the peripheral clocks with the master SCN clock and leads to a metabolic syndrome. Proc. Natl. Acad. Sci. U. S. A. 112, 6691–6698 (2015). 80. Vollmers, C. et al. Time of feeding and the intrinsic circadian clock drive rhythms in hepatic gene expression. Proc. Natl. Acad. Sci. U. S. A. 106, 21453–21458 (2009). 81. Chavan, R. et al. Liver-derived ketone bodies are necessary for food anticipation. Nat. Commun. 7, 1–10 (2016). 82. Mukherji, A., Kobiita, A. & Chambon, P. Shifting the feeding of mice to the rest phase

creates metabolic alterations, which, on their own, shift the peripheral circadian clocks by 12 hours. Proc. Natl. Acad. Sci. U. S. A. 112, 6683–6690 (2015). 83. Schmutz, I., Ripperger, J. A., Baeriswyl-Aebischer, S. & Albrecht, U. The mammalian clock component PERIOD2 coordinates circadian output by interaction with nuclear receptors. Genes Dev. 24, 345–357 (2010). 84. Kornmann, B., Schaad, O., Bujard, H., Takahashi, J. S. & Schibler, U. System-driven and oscillator-dependent circadian transcription in mice with a conditionally active liver clock. PLoS Biol. 5, 0179–0189 (2007). 85. Pan, X. & Hussain, M. M. Clock is important for food and circadian regulation of macronutrient absorption in mice. J. Lipid Res. 50, 1800–1813 (2009). 86. Zhang, R., Lahens, N. F., Ballance, H. I., Hughes, M. E. & Hogenesch, J. B. A circadian gene expression atlas in mammals: implications for biology and medicine. Proc. Natl. Acad. Sci. U. S. A. 111, 16219–24 (2014). 87. Marcheva, B. et al. Disruption of the clock components CLOCK and BMAL1 leads to hypoinsulinaemia and diabetes. Nature 466, 627–631 (2010). 88. Perelis, M. et al. Pancreatic b cell enhancers regulate rhythmic transcription of genes controlling insulin secretion. Science (80-. ). 350, (2015). 89. Kondratov, R. V., Kondratova, A. a., Gorbacheva, V. Y., Vykhovanets, O. V. & Antoch, M. P. Early aging and age-related pathologies in mice deficient in BMAL1, the core component of the circadian clock. Genes Dev. 20, 1868–1873 (2006). 90. Yang, G. et al. Timing of expression of the core clock gene Bmal1 influences its effects on aging and survival. Sci. Transl. Med. 8, 324ra16-324ra16 (2016). 91. Moore, S. R. et al. Robust circadian rhythms in organoid cultures from PERIOD2::LUCIFERASE mouse small intestine. Dis. Model. Mech. 1123–1130 (2014). 92. Matsu-ura, T. et al. Intercellular Coupling of the Cell Cycle and Circadian Clock in Adult Stem Cell Culture. Mol. Cell 0, 237–246 (2016). 93. Barker, N. et al. Identification of stem cells in small intestine and colon by marker gene Lgr5. Nature 449, 1003–1007 (2007). 94. Kim, K. A. et al. Medicine: Mitogenic influence of human R-spondin1 on the intestinal epithelium. Science (80-. ). 309, 1256–1259 (2005). 95. Sato, T. et al. Long-term expansion of epithelial organoids from human colon, adenoma, adenocarcinoma, and Barrett’s epithelium. Gastroenterology 141, 1762–1772 (2011). 96. Clevers, H., Loh, K. M. & Nusse, R. An integral program for tissue renewal and regeneration: Wnt signaling and stem cell control. Science 346, 1248012 (2014). 97. Clevers, H. The intestinal crypt, a prototype stem cell compartment. Cell 154, 274–284 (2013). 98. van der Flier, L. G. & Clevers, H. Stem Cells, Self-Renewal, and Differentiation in the Intestinal Epithelium. Annu. Rev. Physiol. 71, 241–260 (2009). 99. Snippert, H. J. et al. Intestinal crypt homeostasis results from neutral competition between symmetrically dividing Lgr5 stem cells. Cell 143, 134–144 (2010). 100. Sato, T. et al. Single Lgr5 stem cells build crypt-villus structures in vitro without a mesenchymal niche. Nature 459, 262–5 (2009). 101. Mahe, M. M., Sundaram, N., Watson, C. L., Shroyer, N. F. & Helmrath, M. a. Establishment of Human Epithelial Enteroids and Colonoids from Whole Tissue and Biopsy. J. Vis. Exp. 1–13 (2015). doi:10.3791/52483 102. Gracz, A. D., Ramalingam, S. & Magness, S. T. Sox9 expression marks a subset of CD24-expressing small intestine epithelial stem cells that form organoids in vitro. Am. J. Physiol. - Gastrointest. Liver Physiol. 298, 590–600 (2010). 103. Spence, J. R. et al. Directed differentiation of human pluripotent stem cells into intestinal tissue in vitro. Nature 470, 105–9 (2011). 104. Wells, J. M. & Spence, J. R. How to make an intestine. Development 141, 752–760

(2014). 105. Leslie, J. L. et al. Persistence and toxin production by Clostridium difficile within human intestinal organoids result in disruption of epithelial paracellular barrier function. Infect. Immun. 83, 138–145 (2015). 106. Forbester, J. L. et al. Interaction of Salmonella enterica Serovar Typhimurium with Intestinal Organoids Derived from Human Induced Pluripotent Stem Cells. Infect. Immun. 83, 2926–2934 (2015). 107. Sinagoga, K. L. et al. Deriving functional human enteroendocrine cells from pluripotent stem cells. Development 145, dev165795 (2018). 108. Hill, D. R. et al. Bacterial colonization stimulates a complex physiological response in the immature human intestinal epithelium. Elife 6, (2017). 109. Finkbeiner, S. R. et al. Transcriptome-wide Analysis Reveals Hallmarks of Human Intestine Development and Maturation In Vitro and In Vivo. Stem Cell Reports 4, 1140– 1155 (2015). 110. Watson, C. L. et al. An in vivo model of human small intestine using pluripotent stem cells. Nat. Med. 20, 1310–4 (2014). 111. Stokes, K. et al. The Circadian Clock Gene BMAL1 Coordinates Intestinal Regeneration. Cell. Mol. Gastroenterol. Hepatol. 310, F454–F455 (2017). 112. Umemura, Y. et al. Involvement of posttranscriptional regulation of Clock in the emergence of circadian clock oscillation during mouse development. Proc. Natl. Acad. Sci. U. S. A. 114, E7479–E7488 (2017). 113. Bellet, M. M. et al. Circadian clock regulates the host response to Salmonella. Proc. Natl. Acad. Sci. U. S. A. 110, 9897–902 (2013). 114. Zietek, T., Rath, E., Haller, D. & Daniel, H. Intestinal organoids for assessing nutrient transport, sensing and incretin secretion. Sci. Rep. 5, 16831 (2015). 115. Hussain, M. M. & Pan, X. Circadian Regulation of Macronutrient Absorption. J. Biol. Rhythms 30, 459–469 (2015). 116. Sinagoga, K. L. & Wells, J. M. Generating human intestinal tissues from pluripotent stem cells to study development and disease. EMBO J. 34, 1149–63 (2015). 117. Dierickx, P. et al. Circadian networks in human embryonic stem cell‐derived cardiomyocytes. EMBO Rep. 18, 1199–1212 (2017). 118. Edgar, R. S. et al. Cell autonomous regulation of herpes and influenza virus infection by the circadian clock. Proc. Natl. Acad. Sci. 113, 10085–10090 (2016). 119. Kiessling, S. et al. The circadian clock in immune cells controls the magnitude of Leishmania parasite infection. Sci. Rep. 7, 1–11 (2017). 120. Best, E. L., Freeman, J. & Wilcox, M. H. Models for the study of Clostridium difficile infection. Gut Microbes 3, 145–167 (2012). 121. Zhang, Y. G., Wu, S., Xia, Y. & Sun, J. Salmonella-infected crypt-derived intestinal organoid culture system for host–bacterial interactions. Physiol. Rep. 2, 1–11 (2014). 122. Lessa, F. C. et al. Burden of Clostridium difficile Infection in the United States. N. Engl. J. Med. 372, 825–834 (2015). 123. Buffie, C. G. et al. Profound alterations of intestinal microbiota following a single dose of clindamycin results in sustained susceptibility to Clostridium difficile-induced colitis. Infect. Immun. 80, 62–73 (2012). 124. Kelly, C. P., Pothoulakis, C. & LaMont, J. T. Clostridium difficile Colitis. N. Engl. J. Med. 330, 257–262 (1994). 125. Deakin, L. J. et al. The Clostridium difficile spo0A gene is a persistence and transmission factor. Infect. Immun. 80, 2704–2711 (2012). 126. Heap, J. T., Pennington, O. J., Cartman, S. T., Carter, G. P. & Minton, N. P. The ClosTron: A universal gene knock-out system for the genus Clostridium. J. Microbiol. Methods 70, 452–464 (2007). 87

127. Underwood, S. et al. Characterization of the sporulation initiation pathway of Clostridium difficile and its role in toxin production. J. Bacteriol. 191, 7296–7305 (2009). 128. Saujet, L., Monot, M., Dupuy, B., Soutourina, O. & Martin-Verstraete, I. The key sigma factor of transition phase, sigh, controls sporulation, metabolism, and virulence factor expression in clostridium difficile. J. Bacteriol. 193, 3186–3196 (2011). 129. Pettit, L. J. et al. Functional genomics reveals that Clostridium difficile Spo0A coordinates sporulation, virulence and metabolism. BMC Genomics 15, 1–15 (2014). 130. Saujet, L., Pereira, F. C., Henriques, A. O. & Martin-Verstraete, I. The regulatory network controlling spore formation in Clostridium difficile. FEMS Microbiol. Lett. 358, 1–10 (2014). 131. Theriot, C. M. et al. Antibiotic-induced shifts in the mouse gut microbiome and metabolome increase susceptibility to Clostridium difficile infection. Nat. Commun. 5, (2014). 132. Buffie, C. G. et al. Precision microbiome reconstitution restores bile acid mediated resistance to Clostridium difficile. Nature 517, 205–208 (2015). 133. Theriot, C. M., Bowman, A. A. & Young, B. difficile Spore Germination and Outgrowth in the Large Intestine. Am. Soc. Microbiol. 1, 1–16 (2016). 134. Martin-Verstraete, I., Peltier, J. & Dupuy, B. The regulatory networks that control Clostridium difficile toxin synthesis. Toxins (Basel). 8, 1–24 (2016). 135. Mani, N. & Dupuy, B. Regulation of toxin synthesis in Clostridium difficile by an alternative RNA polymerase sigma factor. Proc. Natl. Acad. Sci. U. S. A. 98, 5844–5849 (2001). 136. Matamouros, S., England, P. & Dupuy, B. Clostridium difficile toxin expression is inhibited by the novel regulator TcdC. Mol. Microbiol. 64, 1274–1288 (2007). 137. MacCannell, D. R. et al. Molecular analysis of Clostridium difficile PCR ribotype 027 isolates from Eastern and Western Canada. J. Clin. Microbiol. 44, 2147–2152 (2006). 138. Govind, R. & Dupuy, B. Secretion of Clostridium difficile toxins A and B requires the holin- like protein tcdE. PLoS Pathog. 8, 1–14 (2012). 139. Yuan, P. et al. Chondroitin sulfate proteoglycan 4 functions as the cellular receptor for Clostridium difficile toxin B. Cell Res. 25, 157–168 (2015). 140. Tao, L. et al. Frizzled proteins are colonic epithelial receptors for C. difficile toxin B. Nature 538, 350–355 (2016). 141. LaFrance, M. E. et al. Identification of an epithelial cell receptor responsible for Clostridium difficile TcdB-induced cytotoxicity. Proc. Natl. Acad. Sci. 112, 7073–7078 (2015). 142. Chandrasekaran, R. & Lacy, D. B. The role of toxins in clostridium difficile infection. FEMS Microbiol. Rev. 41, 723–750 (2017). 143. Papatheodorou, P., Zamboglou, C., Genisyuerek, S., Guttenberg, G. & Aktories, K. Clostridial glucosylating toxins enter cells via clathrin-mediated endocytosis. PLoS One 5, e10673 (2010). 144. Henriques, B., Florin, I. & Thelestam, M. Cellular internalisation of Clostridium difficile toxin A. Microb. Pathog. 2, 455–463 (1987). 145. Qa’dan, M., Spyres, L. M. & Ballard, J. D. pH-induced conformational changes in Clostridium difficile toxin B. Infect. Immun. 68, 2470–2474 (2000). 146. Reineke, J. et al. Autocatalytic cleavage of Clostridium difficile toxin B. Nature 446, 415– 419 (2007). 147. Egerer, M., Giesemann, T., Jank, T., Fullner Satchell, K. J. & Aktories, K. Auto-catalytic cleavage of Clostridium difficile toxins A and B depends on cysteine protease activity. J. Biol. Chem. 282, 25314–25321 (2007). 148. Just, I. et al. Glucosylation of Rho proteins by Clostridium difficile toxin B. Nature 375, 500–503 (1995). 149. Hall, A. Rho GTpases and the actin cytoskeleton. Science (80-. ). 279, 509–514 (1998).

150. Gerhard, R. et al. Glucosylation of Rho GTPases by Clostridium difficile toxin A triggers apoptosis in intestinal epithelial cells. J. Med. Microbiol. 57, 765–770 (2008). 151. Chumbler, N. M. et al. Clostridium difficile Toxin B Causes Epithelial Cell Necrosis through an Autoprocessing-Independent Mechanism. PLoS Pathog. 8, (2012). 152. Chumbler, N. M., Farrow, M. A., Lapierre, L. A., Franklin, J. L. & Lacy, D. B. Clostridium difficile toxins TcdA and TcdB cause colonic tissue damage by distinct mechanisms. Infect. Immun. 84, 2871–2877 (2016). 153. Farrow, M. A. et al. Clostridium difficile toxin B-induced necrosis is mediated by the host epithelial cell NADPH oxidase complex. Proc. Natl. Acad. Sci. 110, 18674–18679 (2013). 154. Chan, H. et al. Clostridium difficile toxin B induces autophagic cell death in colonocytes. J. Cell. Mol. Med. 22, 2469–2477 (2018). 155. Lyras, D. et al. Toxin B is essential for virulence of Clostridium difficile. Nature 458, 1176– 9 (2009). 156. Kuehne, S. A., Cartman, S. T. & Minton, N. P. Both, toxin A and toxin B, are important in Clostridium difficile infection. Gut Microbes 2, 711–713 (2011). 157. Drudy, D., Fanning, S. & Kyne, L. Toxin A-negative, toxin B-positive Clostridium difficile. Int. J. Infect. Dis. 11, 5–10 (2007). 158. Lin, Q. et al. Toxin A–Predominant Pathogenic Clostridioides difficile: A Novel Clinical Phenotype. Clin. Infect. Dis. ciz727, 1–6 (2019). 159. Lopez-Urena, D. et al. Toxin B Variants from Clostridium difficile Strains VPI 10463 and NAP1/027 Share Similar Substrate Profile and Cellular Intoxication Kinetics but Use Different Host Cell Entry Factors. Toxins (Basel). 11, 1–15 (2019). 160. Crosby, P. et al. Insulin/IGF-1 Drives PERIOD Synthesis to Entrain Circadian Rhythms with Feeding Time. Cell 177, 896–909 (2019). 161. Reppert, S. M. & Schwartz, W. J. Maternal coordination of the fetal biological clock in utero. Science (80-. ). 220, 969–971 (1983). 162. Stephan, F. K. & Zucker, I. Circadian Rhythms in Drinking Behavior and Locomotor Activity of Rats Are Eliminated by Hypothalamic Lesions (suprachiasmatic and medial preoptic nuclei/retino-hypothalamic projection). Proc.Natl.Acad.Sci. 69, 1583–1586 (1972). 163. Moore, R. Y. & Eichler, V. B. Loss of a circadian adrenal corticosterone rhythm following suprachiasmatic lesions in the rat. Brain Res. Bull. 42, 201–206 (1972). 164. Reppert, S. M. & Schwartz, W. J. Maternal suprachiasmatic nuclei are necessary for maternal coordination of the developing circadian system. J. Neurosci. 6, 2724–2729 (1986). 165. Deguchi, T. Ontogenesis of circadian rhythm of melatonin synthesis in pineal gland of rat. J. Neural Transm. Suppl. 13, 115–128 (1978). 166. Yoo, S.-H. et al. PERIOD2::LUCIFERASE real-time reporting of circadian dynamics reveals persistent circadian oscillations in mouse peripheral tissues. Proc. Natl. Acad. Sci. U. S. A. 101, 5339–5346 (2004). 167. Welz, P. S. et al. BMAL1-Driven Tissue Clocks Respond Independently to Light to Maintain Homeostasis. Cell 177, 1436–1447 (2019). 168. Jud, C. & Albrecht, U. Circadian rhythms in murine pups develop in absence of a functional maternal circadian clock. J. Biol. Rhythms 21, 149–154 (2006). 169. Pandi-Perumal, S. R. et al. Dim light melatonin onset (DLMO): A tool for the analysis of circadian phase in human sleep and chronobiological disorders. Prog. Neuro- Psychopharmacology Biol. Psychiatry 31, 1–11 (2007). 170. McIntyre, I. M., Norman, T. R., Burrows, G. D. & Armstrong, S. M. GUANTAL MELATONIN SUPPRESSION BY EXPOSURE TO LOW INTENSITY LIGHT IN MAN. Life Sci. 45, 327–332 (1989). 171. Ohta, H., Yamazaki, S. & McMahon, D. G. Constant light desynchronizes mammalian

clock neurons. Nat. Neurosci. 8, 267–269 (2005). 172. Houdek, P. et al. Melatonin administered during the fetal stage affects circadian clock in the suprachiasmatic nucleus but not in the liver. Dev. Neurobiol. 75, 131–144 (2015). 173. Yamaguchi, Y. Arginine vasopressin signaling in the suprachiasmatic nucleus on the resilience of circadian clock to jet lag. Neurosci. Res. 129, 57–61 (2018). 174. Canaple, L., Gréchez-Cassiau, A., Delaunay, F., Dkhissi-Benyahya, O. & Samarut, J. Maternal eating behavior is a major synchronizer of fetal and postnatal peripheral clocks in mice. Cell. Mol. Life Sci. 75, 3991–4005 (2018). 175. Sládek, M. et al. Insight into molecular core clock mechanism of embryonic and early postnatal rat suprachiasmatic nucleus. Proc. Natl. Acad. Sci. U. S. A. 101, 6231–6236 (2004). 176. Sládek, M., Jindráková, Z., Bendová, Z. & Sumová, A. Postnatal ontogenesis of the circadian clock within the rat liver. Am. J. Physiol. - Regul. Integr. Comp. Physiol. 292, 1224–1229 (2007). 177. Polidarová, L. et al. Development and entrainment of the colonic circadian clock during ontogenesis. Am. J. Physiol. - Gastrointest. Liver Physiol. 306, 346–356 (2014). 178. Khademhosseini, A. & Langer, R. A decade of progress in tissue engineering. Nat. Protoc. 11, 1775–1781 (2016). 179. Kowalska, E., Moriggi, E., Bauer, C., Dibner, C. & Brown, S. A. The circadian clock starts ticking at a developmentally early stage. J. Biol. Rhythms 25, 442–449 (2010). 180. Yagita, K. et al. Development of the circadian oscillator during differentiation of mouse embryonic stem cells in vitro. Proc. Natl. Acad. Sci. 107, 3846–3851 (2010). 181. Umemura, Y. et al. Transcriptional program of Kpna2/Importin-α2 regulates cellular differentiation-coupled circadian clock development in mammalian cells. Proc. Natl. Acad. Sci. U. S. A. 111, E5039–E5048 (2014). 182. Brown, S. A. et al. The period length of fibroblast circadian gene expression varies widely among human individuals. PLoS Biol. 3, (2005). 183. Zou, Winnie, Y. et al. Human Intestinal Enteroids: New Models to Study Gastrointestinal Virus Infections. Methods Mol. Biol. 209, 359–363 (2017). 184. Koo, B. K. et al. Controlled gene expression in primary Lgr5 organoid cultures. Nat. Methods 9, 81–83 (2012). 185. Hughes, M. E. et al. Guidelines for Genome-Scale Analysis of Biological Rhythms. J. Biol. Rhythms 32, 380–393 (2017). 186. Wu, G., Anafi, R. C., Hughes, M. E., Kornacker, K. & Hogenesch, J. B. MetaCycle: An integrated R package to evaluate periodicity in large scale data. Bioinformatics 32, 3351– 3353 (2016). 187. Bugge, A. et al. Rev-Erbα and Rev-Erbβ Protect the Circadian Clock and Metabolic Function. Genes Dev. 26, 657–667 (2012). 188. Colbert, L. H. et al. Exercise Effects on Tumorigenesis in a p53-deficient Mouse Model of Breast Cancer. Med. Sci. Sport. excercise 41, 1597–1605 (2009). 189. Johnson, B. P. et al. Hepatocyte circadian clock controls acetaminophen bioactivation through NADPH-cytochrome P450 oxidoreductase. Proc. Natl. Acad. Sci. 111, 18757– 18762 (2014). 190. Umemura, Y., Maki, I., Tsuchiya, Y., Koike, N. & Yagita, K. Human Circadian Molecular Oscillation Development Using Induced Pluripotent Stem Cells. J. Biol. Rhythms 34, 525– 532 (2019). 191. Kumar, D., Wingate, D. & Ruckebusch, Y. Circadian variation in the propagation velocity of the migrating motor complex. Gastroenterology 91, 926–930 (1986). 192. Froy, O. & Chapnik, N. Circadian oscillation of innate immunity components in mouse small intestine. Mol. Immunol. 44, 1954–1960 (2007). 193. Schibler, U. et al. Clock-Talk: Interactions between central and peripheral circadian

oscillators in mammals. Cold Spring Harb. Symp. Quant. Biol. 80, 223–232 (2015). 194. Thaiss, C. A. et al. Article Transkingdom Control of Microbiota Diurnal Oscillations Promotes Metabolic Homeostasis. Cell 159, 514–529 (2014). 195. Mukherji, A., Kobiita, A., Ye, T. & Chambon, P. Homeostasis in intestinal epithelium is orchestrated by the circadian clock and microbiota cues transduced by TLRs. Cell 153, 812–827 (2013). 196. Thaiss, C. A. et al. Microbiota Diurnal Rhythmicity Programs Host Transcriptome Oscillations. Cell 167, 1495-1510.e12 (2016). 197. Chandrasekaran, R., Kenworthy, A. K. & Lacy, D. B. Clostridium difficile Toxin A Undergoes Clathrin-Independent, PACSIN2-Dependent Endocytosis. PLoS Pathog. 12, 1–30 (2016). 198. Abt, M. C., McKenney, P. T. & Pamer, E. G. Clostridium difficile colitis: Pathogenesis and host defence. Nat. Rev. Microbiol. 14, 609–620 (2016). 199. Saxena, K. et al. Human Intestinal Enteroids: a New Model To Study Human Rotavirus Infection, Host Restriction, and Pathophysiology. J. Virol. 90, 43–56 (2015). 200. Noel, G. et al. A primary human macrophage-enteroid co-culture model to investigate mucosal gut physiology and host-pathogen interactions. Sci. Rep. 7, 1–14 (2017). 201. Nalapareddy, K. et al. Canonical Wnt Signaling Ameliorates Aging of Intestinal Stem Cells. Cell Rep. 18, 2608–2621 (2017). 202. Bray, N. L., Pimentel, H., Melsted, P. & Pachter, L. Near-optimal probabilistic RNA-seq quantification. Nat. Biotechnol. 34, 525–527 (2016). 203. Emig, D. et al. AltAnalyze and DomainGraph: Analyzing and visualizing exon expression data. Nucleic Acids Res. 38, 755–762 (2010). 204. Balsalobre, A. et al. Resetting of circadian time in peripheral tissues by glucocorticoid signaling. Science (80-. ). 289, 2344–2347 (2000). 205. Hughes, M. E. et al. Harmonics of circadian gene transcription in mammals. PLoS Genet. 5, (2009). 206. Wu, G. et al. Population level rhythms in human skin : implications for circadian medicine Affiliations : Proc. Natl. Acad. Sci. 1–6 (2018). doi:10.1073/pnas.1809442115 207. Saavedra, P. H. V. et al. Apoptosis of intestinal epithelial cells restricts Clostridium difficile infection in a model of pseudomembranous colitis. Nat. Commun. 9, 1–10 (2018). 208. Chang, H.-C. & Guarente, L. SIRT1 Mediates Central Circadian Control in the SCN by a Mechanism that Decays with Aging. Cell 153, 1448–1460 (2013). 209. Sato, S. et al. Circadian Reprogramming in the Liver Identifies Metabolic Pathways of Aging. Cell 170, 664-677.e11 (2017). 210. Zhdanova, I. V et al. Aging of intrinsic circadian rhythms and sleep in a diurnal nonhuman primate, Macaca mulatta. J. Biol. Rhythms 26, 149–159 (2011). 211. Davidson, a. J. et al. Chronic jet-lag increases mortality in aged mice. Curr. Biol. 16, 7– 10 (2006). 212. Pechal, A., Lin, K., Allen, S. & Reveles, K. National age group trends in Clostridium difficile infection incidence and health outcomes in United States Community Hospitals. BMC Infect. Dis. 16, 1–6 (2016). 213. Kiessling, S. et al. Enhancing circadian clock function in cancer cells inhibits tumor growth. BMC Biol. 15, 1–18 (2017). 214. Chen, R., Seo, D. O., Bell, E., Von Gall, C. & Lee, D. C. Strong resetting of the mammalian clock by constant light followed by constant darkness. J. Neurosci. 28, 11839–11847 (2008). 215. Polidarová, L., Sládek, M., Soták, M., Pácha, J. & Sumová, A. Hepatic, duodenal, and colonic circadian clocks differ in their persistence under conditions of constant light and in their entrainment by restricted feeding. Chronobiol. Int. 28, 204–215 (2011). 216. Mori-Akiyama, Y. et al. SOX9 Is Required for the Differentiation of Paneth Cells in the

Intestinal Epithelium. Gastroenterology 133, 539–546 (2007). 217. Farin, H. F., Van Es, J. H. & Clevers, H. Redundant sources of Wnt regulate intestinal stem cells and promote formation of paneth cells. Gastroenterology 143, 1518–1529 (2012). 218. Van Opstal, E. et al. Vancomycin Treatment Alters Humoral Immunity and Intestinal Microbiota in an Aged Mouse Model of Clostridium difficile Infection. J. Infect. Dis. 214, 130–139 (2016). 219. Shin, J. H. et al. Innate Immune Response and Outcome of Clostridium difficile Infection Are Dependent on Fecal Bacterial Composition in the Aged Host. J. Infect. Dis. 217, 188– 197 (2018). 220. Nakahata, Y., Sahar, S., Astarita, G., Kaluzova, M. & Sassone-Corsi, P. Circadian control of the NAD+ salvage pathway by CLOCK-SIRT1. Science 324, 654–657 (2009). 221. Ramsey, K. M. et al. Circadian Clock Feedback Cycle Through NAMPT-Mediated NAD+ Biosynthesis. Science (80-. ). 324, 651–654 (2009). 222. Nakahata, Y. et al. The NAD+-Dependent Deacetylase SIRT1 Modulates CLOCK- Mediated Chromatin Remodeling and Circadian Control. Cell 134, 329–340 (2008). 223. Asher, G. et al. SIRT1 Regulates Circadian Clock Gene Expression through PER2 Deacetylation. Cell 134, 317–28 (2008). 224. Lin, S. J., Defossez, P. a & Guarente, L. Requirement of NAD and SIR2 for life-span extension by calorie restriction in Saccharomyces cerevisiae. Science 289, 2126–2128 (2000). 225. Rajman, L., Chwalek, K. & Sinclair, D. A. Therapeutic Potential of NAD-Boosting Molecules: The In Vivo Evidence. Cell Metab. 27, 529–547 (2018). 226. Yang, H. et al. Nutrient-sensitive mitochondrial NAD+ levels dictate cell survival. Cell 130, 1095–1107 (2007). 227. Berger, F., Lau, C., Dahlmann, M. & Ziegler, M. Subcellular compartmentation and differential catalytic properties of the three human nicotinamide mononucleotide adenylyltransferase isoforms. J. Biol. Chem. 280, 36334–36341 (2005). 228. Mills, K. F. et al. Long-Term Administration of Nicotinamide Mononucleotide Mitigates Age-Associated Physiological Decline in Mice. Cell Metab. 24, 795–806 (2016). 229. Pajk, M. et al. Exogenous nicotinamide supplementation and moderate physical exercise can attenuate the aging process in skeletal muscle of rats. Biogerontology 18, 593–600 (2017). 230. Bitterman, K. J., Anderson, R. M., Cohen, H. Y., Latorre-Esteves, M. & Sinclair, D. A. Inhibition of silencing and accelerated aging by nicotinamide, a putative negative regulator of yeast Sir2 and human SIRT1. J. Biol. Chem. 277, 45099–45107 (2002). 231. Park, S. J. et al. Resveratrol ameliorates aging-related metabolic phenotypes by inhibiting cAMP phosphodiesterases. Cell 148, 421–433 (2012). 232. Mitchell, S. J. et al. The SIRT1 activator SRT1720 extends lifespan and improves health of mice fed a standard diet. Cell Rep. 6, 836–843 (2014). 233. He, B. et al. The Small Molecule Nobiletin Targets the Molecular Oscillator to Enhance Circadian Rhythms and Protect against Metabolic Syndrome. Cell Metab. 23, 610–621 (2016). 234. Nohara, K. et al. Nobiletin fortifies mitochondrial respiration in skeletal muscle to promote healthy aging against metabolic challenge. Nat. Commun. 10, 1–15 (2019). 235. Gertz, M. et al. Ex-527 inhibits Sirtuins by exploiting their unique NAD+-dependent deacetylation mechanism. Proc. Natl. Acad. Sci. U. S. A. 110, (2013). 236. Hasmann, M. & Schemainda, I. FK866, a Highly Specific Noncompetitive Inhibitor of Nicotinamide Phosphoribosyltransferase, Represents a Novel Mechanism for Induction of Tumor Cell Apoptosis. Cancer Res. 63, 7436–7442 (2003). 237. Voigt, R. M. et al. Circadian disorganization alters intestinal microbiota. PLoS One 9,

(2014).

Appendix:

R.script for running FFT analysis of rhythmic bioluminescence signal for quantification of amplitude and period length.

# This version of template does the period analysis by Fast Fourier Transform (FFT) with demeaned, detrended, tapered, repeated and then padded data.

# To run this program, please make sure that you have already successfully installed the package

"nnet". If not, please install it first. In the R window, choose "Packages", then choose "Install package(s)...", then choose a nearest mirror site, for example "USA (OH)", click OK. Scroll down to choose "nnet" package to intall.

# It is assumed your input data is saved as "*.csv" with tab delimited. In the data, each column records the observations for one gene, and each row records one time point. There is a header row for gene names.

# You need to specify several things to run the program:

# 1.DIR - the working directory (path of the folder containing your input data and to save your output result);

# 2.DATAFILE - data file name;

# 3.START, END - the time point locations of the first and last peaks for each gene;

# 4.N.Genes - number of genes;

# 5.LAG - the time lag between consecutive time points.

# Example:

# 1. If your data is saved in the folder C:\Users\Your name\Documents, then your DIR should be

"C:\\Users\\Your name\\Documents".

# 2. If your data is saved as the name data.csv, then your DATAFILE should be "data.csv".

# 3. If your data contain 5 genes, and the 5 genes achieve the first peak at time 3, 2, 5, 3, and 3, respectively, then your START value is c(3,2,5,3,3). Suppose also that the 5 genes acheive the last peak at time 128, 130, 126, 133, and 130, respectively, then your END value is c(128,130,126,133,130).

# 4. For this example, your N.Genes value is 5.

# 5. If your first time point is Hour0, and your second time point is Hour4, then your LAG value is

# Specify your values

DIR="C:\\Users\\XX\\Desktop\\FFT"

DATAFILE="1 - Period analysis file.csv"

START=c()

END=c()

N.Genes=

LAG=

# Include package

library(nnet)

# Set the working directory

setwd(DIR)

# Read in data

data0=read.csv(DATAFILE,header=T) data=t(matrix(unlist(data0),ncol=N.Genes))

# Prepare output results

phase=array(rep(NA,N.Genes),dim=c(N.Genes,1),dimnames=list(dimnames(data0)[[2]],c("phas e"))) period=array(rep(NA,N.Genes),dim=c(N.Genes,1),dimnames=list(dimnames(data0)[[2]],c("perio d"))) amplitude=array(rep(NA,N.Genes),dim=c(N.Genes,1),dimnames=list(dimnames(data0)[[2]],c("a mplitude")))

# FFT time length table

N.T=ncol(data) base=NULL for (i in 0:floor(log(N.T,2))) {

for (j in 0:floor(log(N.T,3))) {

for (k in 0:floor(log(N.T,5))) {

base=cbind(base, 2^i*3^j*5^k)

}

base=sort(as.vector(base))[2:(sum(base<=N.T)+1)]

# Start to compute

for (i in 1:N.Genes) {

data1=data[i,START[i]:END[i]]

d=data1-mean(data1)

t=1:length(data1)

res=lm(d ~ t)$resid # If you think your data have a quadratic trend, then replace this statement by "res=lm(d ~ I(t^2)+ t)$resid"

taper=spec.taper(res,p=0.1)

N.TimeSeries=base[(base>=length(data1))][1]

series_rep=as.vector(c(rep(taper,10),rep(0,(N.TimeSeries-length(data1))*10)))

f=1:(length(series_rep)/2)/length(series_rep)

x_label=rev(round(LAG/f,1))

fourier=fft(series_rep)

phase.all=atan(Im(fourier)[2:(N.TimeSeries*5+1)]/Re(fourier)[2:(N.TimeSeries*5+1)])

P=(abs(fourier)^2*4/(N.TimeSeries^2*100))[2:(N.TimeSeries*5+1)]

period[i,]=round(LAG*length(series_rep)/which.is.max(P),2)

amplitude[i,]=round(sqrt(P[which.is.max(P)]),2)

phase[i,]=round(phase.all[which.is.max(P)]/(2*pi)*period[i,],2)

png(paste("plot for gene ",dimnames(data0)[[2]][i],".png",sep=""),width=850,height=600)

barplot(rev(P),xlab="Period

(Hour)",ylab="Periodogram",names.arg=x_label,main=paste("Periodogram (Period=",period[i,],",

Amplitude=",amplitude[i,],",

Phase=",phase[i,],")",sep=""),cex.main=2,cex.axis=2,cex.lab=2,cex.names=2)

dev.off()

x_label2=rev(LAG/f)

spec.out=cbind(x_label2[x_label2>=5 & x_label2<=20],rev(P)[x_label2>=5 & x_label2<=20])

colnames(spec.out)=c("Period","Spectrum")

write.csv(spec.out,paste("spec.out_",i,".csv",sep=""))

}

write.csv(phase,"phase.csv") write.csv(period,"period.csv") write.csv(amplitude,"amplitude.csv")