Modification of Neurogenic Colonic Motor Behaviours by Chemogenetic Ablation of Calretinin Neurons

How the enteric nervous system determines the pacing and propagation direction of neurogenic contractions along the colon remains largely unknown. We used a chemogenetic strategy to ablate enteric neurons expressing calretinin (CAL). Mice expressing human diphtheria toxin receptor (DTR) in CAL neurons were generated by crossing CAL-ires-Cre mice with Cre-dependent ROSA26-DTR mice. Immunohistochemical analysis revealed treatment with diphtheria toxin incurred a 42% reduction in counts of Hu-expressing colonic myenteric neurons (P = 0.036), and 57% loss of CAL neurons (comprising ∼25% of all Hu neurons; P = 0.004) compared to control. As proportions of Hu-expressing neurons, CAL neurons that contained nitric oxide synthase (NOS) were relatively spared (control: 15 ± 2%, CAL-DTR: 13 ± 1%; P = 0.145), while calretinin neurons lacking NOS were significantly reduced (control: 26 ± 2%, CAL-DTR: 18 ± 5%; P = 0.010). Colonic length and pellet sizes were significantly reduced without overt inflammation or changes in ganglionic density. Interestingly, colonic motor complexes (CMCs) persisted with increased frequency (mid-colon interval 111 ± 19 vs. 189 ± 24 s, CAL-DTR vs. control, respectively, P < 0.001), decreased contraction size (mid-colon AUC 26 ± 24 vs. 59 ± 13 gram/seconds, CAL-DTR vs. control, respectively, P < 0.001), and lacked preferential anterograde migration (P < 0.001). The functional effects of modest calretinin neuron ablation, particularly increased neurogenic motor activity frequencies, differ from models that incur general enteric neuron loss, and suggest calretinin neurons may contribute to pacing, force, and polarity of CMCs in the large bowel.


INTRODUCTION
In vertebrate large intestine, a number of neurogenic motor patterns occur that require or involve the enteric nervous system Spencer and Hu, 2020). In mouse, the colonic motor complex (CMC) is analogous to the cyclic motor complexes in guinea pig colon. This behaviour comprises regularly occurring neurogenic contractions of the longitudinal and circular muscle of the large intestine that propel content over considerable distances and have been recorded from a variety of species, including humans (Bampton et al., 2000;Hagger et al., 2002;Spencer et al., 2012;, mice (Fida et al., 1997;Balasuriya et al., 2016;Hibberd et al., 2017;Spencer et al., 2018), guinea-pigs (Costa and Furness, 1976;Costa et al., 2013), and rabbits (Fida et al., 1997;Powell et al., 2002;Dinning et al., 2012;Costa et al., 2013). It is well-established that CMCs require enteric neurons for coordination as they are immediately abolished by neuronal blockade with tetrodotoxin (TTX) (Wood et al., 1986;Fida et al., 1997;Brierley et al., 2001;Powell et al., 2002;Spencer and Bywater, 2002; or by blockade of cholinergic nicotinic receptors (Bush et al., 2000;Brierley et al., 2001;Powell et al., 2002;, the primary mode of fast neurotransmission among enteric neurons (Galligan, 2002). All enteric neurons, including those that generate CMCs, can be classified by their neurochemical content. Much is known about the different neurochemical classes of enteric neurons that exist and their relationships to functional classifications (Costa et al., 1996;Young, 1996, 1998;Lomax and Furness, 2000). However, complex motor behaviours and neurochemical classification are typically studied in isolation. Thus, little is known about how neurochemical classes of enteric neurons contribute to complex neurogenic motor behaviours.
A major population of neurons in the large intestine of mice express the neurochemical marker, calretinin which comprises about 30-50% of myenteric neurons (Sang and Young, 1996;Musser et al., 2015). Calretinin neurons comprise myenteric motor neurons, interneurons and importantly, the vast majority of putative primary afferent neurons that also express calcitonin gene related peptide (CGRP) (99%) (Sang and Young, 1996;Sang and Young, 1998;Furness et al., 2004). These are presumptively analogous to the Dogiel type II cells that have been referred to as intrinsic primary afferent neurons (IPANs) and have been shown to be directly mechanosensitive in guinea pig (Kirchgessner et al., 1992;Kunze et al., 1998;Kunze et al., 2000) and in the mouse ileum (Mao et al., 2006). Although, there is also evidence of direct mechanosensitivity among populations of myenteric neurons with Dogiel Type I morphology (i.e., interneurons or motor neurons; Spencer and Smith, 2004;Mazzuoli and Schemann, 2009;Mazzuoli and Schemann, 2012;Mazzuoli-Weber and Schemann, 2015), it has been suggested that IPANs, most of which express calretinin, are the first neurons to activate myenteric neural circuits (Kunze and Furness, 1999).
In this study, we selectively ablated calretinin-expressing neurons in adult mice by expressing the human diphtheria toxin receptor (DTR) in calretinin neurons. This allowed us for the first time to characterise the effects in the large bowel of ablation of calretinin-expressing neurons on motility and in vitro transit in adult mice, without concerns of genetic compensation. The results suggest that enteric calretinin neurons may modulate pacemaker frequency and propagation direction of CMCs.

Animals
Calb2-IRES-Cre [B6(Cg)-Calb2 TM1(cre)Zjh /J] [stock #010774 (Taniguchi et al., 2011)] and Rosa-DTR mice [stock #016603 (Buch et al., 2005)] obtained from Jackson Laboratories (Bar Harbor, ME, United States) were crossbred ( Figure 1A). Resulting CAL cre+ -DTR f /f progeny co-expressed human DTR with the Calb2 gene product, calretinin. All mice were housed under a 12-h light/dark cycle with food and water provided ad libitum. Age-matched, Cre − littermates were used as control in all experiments. Mice ranged 8-12-weeks old in all experiments. Animal studies are reported in compliance with the ARRIVE guidelines, as well as the guidelines of the National Institutes of Health and the International Association for the Study of Pain and were approved by the Animal Studies Committee at Washington University School of Medicine and conform to the principles and regulations as described in the Editorial by Grundy (2015).

Diptheria Toxin Treatment
Depletion of calretinin neurons was accomplished by using intraperitoneal injections of diphtheria toxin (diptheria toxin (DTX); 600 ng per mouse) every 48 h for a total of 3 treatments to both cre − control and cre + CAL-DTR mice (Luo et al., 2018). At least 1 week was allowed between the final DTX treatment and experimentation. Ablation efficiency was confirmed by counting CAL-IR myenteric nerve cell bodies after all experiments.

Tissue Preparation
The abdomen was immediately opened after the mouse was euthanised and the entire colon from caecum to terminal rectum was removed. In all experiments except the imaging of natural pellet expulsion, the full length of colon was placed in a petri dish filled with carbogen-gassed (95% O 2 /5% CO 2 ) Krebs solution (25-30 • C; in 10 −3 M concentrations: NaCl 118; KCl shows an expanded region of mucosa from the dotted square region in panel (F). Note the similarity (i.e., lack of overt inflammation) in mucosa in panel (G) as in control mouse in panel (D). Panel (G) shows circular muscle thicknesses in control and CAL-DTR mice. Panel (H) shows longitudinal muscle thicknesses in control and CAL-DTR mice. Panel (I) shows circular muscle thickness in control and CAL-DTR mice. Panel (J) shows goblet cell densities in control and CAL-DTR mice. Each dot point represents a mean value from each animal studied. A total of N = 5 control and N = 3 DTX mice were used in analysis. There was no significant difference between any of the parameters measured between control and CAL-DTR mice. 4.7, NaH 2 PO 4 1; NaHCO 3 25; MgCl 2 1.2; D-Glucose 11; CaCl 2 2.5). Whole colon length was measured before emptying by gentle flushing with Krebs solution and removal of the mesentery. For recording natural pellet expulsion in vitro, the terminal rectum was occluded by tying off with fine suture thread before the colon was removed from the animal.

Haematoxylin and Eosin
Colons were fixed in 4% paraformaldehyde in phosphatebuffered saline (PBS), dehydrated in 30% sucrose, and embedded in optimal cutting temperature (OCT) compound. Twelvemicron sections were made using a Leica CM1950 cryostat (Leica Biosystems, Buffalo Grove, IL, United States). Haematoxylin and Eosin (H&E) staining was performed according to standard protocols.

Immunohistochemistry
Full-length colons were cleared of content and cut along the mesenteric border. Flat sheet preparations were stretched maximally and fixed for 48 h in paraformaldehyde (4% in 0.1 M phosphate buffer, pH 7.0). The mucosa was removed by sharp dissection and the submucosal layer was peeled off the underlying circular muscle intact. Longitudinal muscle-myenteric plexus preparations were obtained by removing the remaining circular muscle. Preparations were cleared using 0.5% Triton-X100 in 0.1 M phosphate buffered saline (PBS; 0.15 M NaCl, pH 7.2; 3 × 10 min washes and then washed in PBS (3 × 10 min washes) followed by incubation.
Whole length preparations of colonic longitudinal musclemyenteric plexus were incubated for 30 min to 1 h in phosphate buffered saline (PBS) + 0.1% Triton X-100 containing 10% Preparations were then washed with PBS (3 × 10 min) and equilibrated in a series of carbonate-buffered glycerol solutions (50, 70, and 100% solutions; 3 × 10 min) before mounting on glass slides in buffered glycerol (pH 8.6). Preparations were viewed and imaged using an epifluorescence microscope (Olympus IX71, Japan) equipped with discriminating filters to match the fluorophores used (Chroma Technology Co., Battledore, VT). Images were captured by a Roper Scientific camera and AnalySIS Imager 5.0 software (Olympus-SIS, Munster, Germany) via 20 × or 40 × water immersion lenses. Images were stored as TIFF files (1392 * 1080 pixels) and optimised for contrast and brightness using Adobe Photoshop (2021 Adobe Systems Software Ireland Ltd., Mountain View, CA, United States) before analysis. CAL-IR and NOS-IR myenteric nerve cell bodies were quantified by performing counts from a minimum 15 randomly selected fields of view taken as approximately equal intervals along the length of the colon. The field of view (FOV) encompassed a region of 465 × 465 µm. Analyses were performed by identifying all Hu-immunoreactive myenteric nerve cell bodies before assessing their NOS and calretinin immunoreactive content. Assessment was performed blinded to preparation genotype. Analysis was performed on micrographs taken using 20 × immersion lens, comprising a 0.216 mm 2 field of view. Cell counts in CAL-DTR colons were corrected by a factor of 0.757 to account for mean change in colon size. Both corrected and uncorrected counts are presented.

Mechanical and Electrophysiological Recordings
Combined mechanical and electrophysiological recordings were made from full-length colon preparations in an organ bath, similar to that described previously (Hibberd et al., 2018a). The organ bath (volume ∼50 ml) was continuously superfused at ∼5 ml.min −1 (36 • C). A stainless steel tube (diameter 1 mm) placed through the lumen was fitted at each end into L-shaped barbed plastic connectors that were fixed to organ bath base. The oral and anal ends of colon were tied over the barbed connectors with a fine suture thread. Stainless steel hooks (250 µm diameter) were threaded through the wall of the proximal, middle, and distal colon. A fine suture thread connected each hook to an isometric force transducer (MLT050/D; ADInstruments, Bella Vista, NSW, Australia) and set to a basal tension of 0.5 g to record mechanical activity. Activity was recorded at 1 kHz (PowerLab 16/35, LabChart 8, ADInstruments, Bella Vista, NSW, Australia).
During mechanical recordings, conventional extracellular recordings were made from the serosal surface of the colon. Two Krebs-filled suction electrodes (AgCl, 250 µm) with heatpolished glass capillary tips were used (0.86 mm internal diameter, 1.5 mm outer diameter; cat# 30-0053, Harvard Apparatus, Holliston, MA, United States). Flexible Silastic tubing (Dow Corning Corporation, Midland, MI, United States) connected the glass tip to the AgCl electrode. This enabled electrodes to remain in position during gut movement. Signals were amplified by a factor of 1000 (ISO80; WPI, Sarasota, FL, United States) and recorded at 1 kHz (PowerLab 16/35, LabChart 8, ADInstruments, Bella Vista, NSW, Australia). The signals recorded were in principle assumed to represent extracellular recording of changes in polarity along the surface of the smooth muscle cells. Electrodes are referred as "oral" and "aboral" electrodes in Figures 7D-F. However, the positioning of electrodes was similar to that described previously (Hibberd et al., 2018a), whereby both electrodes were located approximately 10 mm apart, between the mid and distal hooks used for mechanical recordings.

Natural Pellet Expulsion and Spatio-Temporal Mapping
The full length of colon containing natural pellets was placed into a Krebs-filled, Sylgard-lined organ bath (Sylgard 184, Dow Corning Corporation, Midland, MI, United States). Preparations were held in place without obstructing the passage of pellets using etymological pins [500 µm diameter (Barnes et al., 2014)]. The suture was then cut free and expulsion of natural pellets was observed over the following 30 min. Gut movements were videographed by a camera placed above the organ bath (Carl Zeiss Tessar C920, Logitech International S.A., Lausanne, Switzerland). Videographs were transformed into maps of circumferential gut diameter (DMaps) using the spatio-temporal mapping technique described by Hennig et al. (1999).

Statistics and Analysis
Spatio-temporal maps were generated from video recordings using computer script for Matlab software (MathWorks Incorporated, Natick, MA, United States). Script was written in-house at Flinders University (Lukasz Wiklendt, Flinders University). DMaps are generated by measuring gut diameter at each point along the preparation in every video frame and converting diameters into grayscale pixels to create a spatiotemporal map of diameter changes (Figure 8). Regions of minimal diameter (contraction) are represented as white pixels, whereas maximal diameter (distension) are represented by black pixels. This method was adapted from what was described in detail elsewhere (Hennig et al., 1999). Diameter maps were analysed manually using PlotHRM software (Lukasz Wiklendt, Flinders University), written in Matlab (Mathworks) and JavaTM (Sun Microsystems, Santa Clara, CA, United States).
Statistical differences were considered significant if P < 0.05. Results are expressed as mean ± standard deviation except where otherwise stated. Lower case "n" always indicates the number of animals used in a set of experiments. Statistical analysis was performed by ANOVA (one-way or two-way), Student's twotailed t-test for paired or unpaired data using Prism 6 (GraphPad Software, Inc., La Jolla, CA, United States). All authors had access to the study data and had reviewed and approved the final manuscript.

Colonic Length, Muscle Thickness, and Inflammation
Whole colons of DTX-treated CAL-DTR and age-matched control mice were removed and measured. Compared to control, colonic length of DTX-treated CAL-DTR mice was significantly reduced (length 6.75 ± 0.51 cm vs. 5.10 ± 0.50 cm, n = 6 and 4, control and CAL-DTR, respectively, P = 0.001, independent samples t-test). To determine whether the colonic mucosa underwent anatomical and architectural changes after DTX treatment, we compared H&E sections between Cre − (control) and Cre + (DTX-treated CAL-DTR) mice. Following DTX treatment, the cell structure and mucosal architecture was normal in both cohorts of animals, with no evidence mucosal ulcerations or necrosis (Figure 1). To determine if DTX-treated CAL-DTR mice had inflammation induced by DTX treatment, we compared the number of goblet cells in the colon, between control and DTX-treated CAL-DTR mice. There was no difference in goblet cell density between control (30 ± 2 per 100 µm 2 ; N = 3) and DTX-treated CAL-DTR colons [27 ± 2 per 100 µm 2 , N = 5; Welch's t-test, using T distribution (DF = 2.5834) (two-tailed), the P-value = 0.24]. We also compared the thickness of the circular and longitudinal muscle cells between both cohorts of animals. Despite the differences in lengths of the colon, the average thickness of the circular smooth muscle was similar in both groups: control (128 ± 8 µm) and DTX-treated CAL-DTR colons (116 ± 6 µm) (P = 0.69; N = 5; Mann Whitney test). The longitudinal muscle was also not different between the two populations of mice: control (37 ± 2 µm) and DTX-treated CAL-DTR colons (32 ± 2 µm) (P = 0.42; N = 5; Mann Whitney test). Likewise, the diameter of the colons was not significantly different (proximal: 2.39 ± 0.22 vs. 2.35 ± 0.18 mm; mid: 2.13 ± 0.36 vs. 1.8 ± 0.18 mm; distal: 2.3 ± 0.25 vs. 2.02 ± 0.27 mm; n = 5 and 9, control and CAL-DTR, respectively, p = 0.116, two way ANOVA).
A conspicuous difference between control and CAL-DTR preparations was the observation that CMC frequency in CAL-DTR preparations was increased compared with controls. CMCs in control preparations had an average interval of 189 ± 24 s (mid-colon; n = 7) while CMCs in CAL-DTR colon occurred every 111 ± 19 s (P < 0.001, Bonferroni post-test, 2-way ANOVA; control: n = 7, CAL-DTR: n = 7; Table 2 and Figure 7A). The strength and duration of CMCs in CAL-DTR colon was reduced compared to control CMCs: peak amplitude, area under the curve and neurogenic spike burst durations were all significantly different (genotype main effects: P = 0.001, <0.001, <0.001, and 0.047, respectively, 2-way ANOVA). These data, including post hoc comparisons are detailed in Table 2 and Figure 7. Expectedly, neurogenic spike bursts, which underlie CMCs, were also more frequent in the CAL-DTR colon than in controls ( Figure 7D).

Calretinin Neuron Ablation
Chemogenetic ablation using targeted diphtheria receptor expression has been used to study immune signalling in the gut (Dunkin et al., 2017;Kang et al., 2019) and inducible diphtheria toxin subunit A has been used to ablate enteric glia (Rao et al., 2017). The latter showing a slightly higher ablation efficacy of 66 and 74% losses of S100β-IR myenteric and intramuscular glial cells in the small and large intestine, respectively. The targeted DTR method to ablate neurons in the ENS has not previously been used and represents an exciting new step forward to probe the functional roles of enteric neurons.
The statistically significant results of the immunohistochemical analysis in the present study were consistent with most expected effects of calretinin ablation. That is, there was a significant reduction in numbers of Hu-IR neurons, indicating losses of myenteric neurons and not just calretinin protein expression. There were also loss of calretinin neurons both as their  proportion of Hu-IR neurons and in their average numbers per FOV along the colon (see Figure 2 and Table 1). Among calretinin neurons the largest effect (61%) was seen among those that lacking NOS, and they showed reductions both as a proportion of Hu-IR neurons and in their numbers per FOV.
Calretinin neurons that contained NOS were reduced in their counts (50%) but not as a proportion of Hu-IR neurons, possibly in part due to the larger effect seen in those lacking NOS. Whether the reduced effect was related to NOS expression is unclear (Sandgren et al., 2003;Lin et al., 2004). Values represent mean ± s.d.; Abbreviations: NSB, neurogenic spike burst; AUC, area under the curve; HPD, half-peak duration; MFR, mean firing rate.
The effect on Hu-IR neuron counts (losses of 28 cells per FOV), represents a 42% reduction, which is similar to the total proportion of CAL-IR neurons in control preparations of ∼41%. However, ablation of CAL-IR neurons was not complete: the average effect size among the CAL-IR populations (average losses of 15 cells per FOV) represented just over half of the reduction in Hu-IR neurons (28 cells per FOV). Notwithstanding neurochemical plasticity (de novo or increased calretinin expression in surviving myenteric neurons), the discrepancy in these effect sizes may represent evidence of an effect of DTX administration on other neurons, either by off-target cre-driven expression of the diphtheria receptor or secondary to a dependence on ablated calretinin neurons. The likelihood of the former possibility cannot be completely ruled out but is allayed by previous characterisation of the cre-driven expression of eYFP in the same Calb2-IRES-Cre mice used in the present study to generate CAL-DTR mice. In these mice, 97% of eYFP expression occurred in calretinin-immunoreactive neurons (Hibberd et al., 2018a). However, a subset of calretinin neurons (likely the morphologically Dogiel type II population that also contain CGRP; Furness et al., 2004) represent the exclusive source of nerve growth factor among mouse colonic myenteric neurons (Petrie et al., 2015). Thus, calretinin neuron ablation may have led to a loss of NGF signalling with secondary effects on other neurons. However, survival of trkA-expressing enteric neurons was not solely dependent on NGF signalling (Petrie et al., 2015), suggesting involvement of other factors (Liu, 2018). Nevertheless, the possibility of off-target neuronal losses cannot be ruled out despite the apparent selectivity of the statistically significant effects for calretinin-expressing neurons, and this should be considered in interpreting results of the present study.

Colonic Motor Complex Contraction Strength
Decreased contraction strength is an expected consequence of calretinin neuron ablation. Smooth muscle contraction of the murine colon during the CMC is predominantly driven by acetylcholine and tachykinins (Bywater et al., 1989;Bush et al., 2000;Brierley et al., 2001). Circular muscle motor neuron nerve terminals that express choline acetyltransferase, vesicular acetylcholine transporter, and substance P also express calretinin the mouse colon (Sang and Young, 1996;Sang and Young, 1998). Thus, decreased strength of contractions observed in CAL-DTR mice may be explicable by loss of excitatory input to the smooth muscle.

Direction of Contraction Propagation
Colonic motor complexes without intraluminal content propulsion are preferentially anterograde but can occur retrogradely, or synchronously along the gut (Powell and Bywater, 2001;Powell et al., 2002;Keating and Spencer, 2010;Copel et al., 2013;Costa et al., 2020). CMCs in the CAL-DTR colon were characterised by a loss of anterograde preference and a preponderance of CMCs that propagated in the retrograde direction. This suggests calretinin enteric neurons may play a role in regulating CMC polarity. There was a significant main effect of genotype (P < 0.001, 2 way ANOVA, N = 7), in addition to multiple post hoc differences (Bonferroni) across different colonic subregions shown on the graph. Peak amplitudes (B) were significantly different across genotypes (genotype main effect, P = 0.001, 2 way ANOVA, N = 7), without significant post hoc differences in colonic subregions. (C) CMC contraction area under the curve was significantly different across genotypes (genotype main effect, P < 0.001, 2 way ANOVA, N = 7) and by post hoc comparison in the proximal and mid colon. (D) Similar to CMCs, neurogenic spike burst intervals were significantly reduced in CAL-DTR colon, compared to controls (genotype main effect, P < 0.001, 2 way ANOVA, N = 7). Neurogenic spike burst durations (E) were significantly different (genotype main effect, P = 0.047, 2 way ANOVA, N = 7), but not by post hoc test. The rate of action potential discharge within neurogenic spike bursts (F) was not significantly different by either genotype main effect or post hoc test. (G) As with CMCs and neurogenic spike bursts, propagating contractions in diameter-mapped preparations showed decreased intervals in CAL-DTR colon, compared to control. Contractions also propagated with higher velocity in CAL-DTR colon, compared to control (H). (I) The maximum cross-sectional area of faecal pellets was substantially reduced in CAL-DTR colons, compared to control. ***P < 0.001, **P < 0.01, *P < 0.05, N = 7. NSB, neurogenic spike burst. FIGURE 8 | Propagating contractions in control and CAL-DTR colon. In these diameter maps, changes in gut diameter are shown with darker areas indicating increases in diameter while lighter areas indicate decreases in diameter. Both preparations contained pellets (control: seven pellets, CAL-DTR: four pellets). Time increases from left to right and the gut is orientated with proximal and distal ends from top to bottom. Control colon (left) shows 2 contractions in a 10 min period, compared to propagating contractions in the CAL-DTR colon (right). Note the higher frequency in CAL-DTR colon, with 6 contractions in the same 10 min period. All contractions arise from the most proximal end of the gut.
Whilst CMCs do not require intraluminal content, movement of intraluminal content may involve an additional mechanism whereby local polarised neural circuits are sequentially activated by content in a neuromechanical cycle (Dinning P. G. et al., 2014;Costa et al., 2015). Indeed, the polarity of intraluminal fluid (Roberts et al., 2007;Roberts et al., 2008;Costa et al., 2013;Sasselli et al., 2013;Wafai et al., 2013;Welch et al., 2014;Barnes and Spencer, 2015;Balasuriya et al., 2016;McQuade et al., 2016;Robinson et al., 2017;Vincent et al., 2018;Israelyan et al., 2019;, and pellet propagation, is virtually always anterograde (Barnes et al., 2014;Balasuriya et al., 2016;Diss et al., 2016;McQuade et al., 2017;Hibberd et al., 2018a;McQuade et al., 2018). In contrast to the loss of anterograde preference in non-propulsive CMCs in CAL-DTR colon in the present study (Figure 6), there was no difference in propagating contraction direction between control and CAL-DTR colon during natural pellet expulsion: nearly all contractions in both groups propagated anterogradely. Thus, the polarity of colonic propulsive motility withstands the losses of calretinin neurons imposed in the present study. Short distances of pellet "retropulsion" have been observed (Brann and Wood, 1976;Heredia et al., 2013), however, we are unaware of any reports of reversal of colonic propulsion in vitro or in vivo, such that pellets located in the distal colon may be propelled and eventually expelled from the proximal colon. We speculate that calretinin neurons contribute more heavily to non-propulsive CMC polarity, than to the robust polarity of propulsive contractions, possibly due to additional recruitment of local polarised circuits. Indeed, normal anterograde propulsion in the gut is fundamental to life and interruption of anterograde propulsion, as occurs with deletion of choline acetyl transferase from the ENS (Johnson et al., 2018) or in aganglionosis (Pichel et al., 1996;Shen et al., 2002), may be lethal.

Colonic Motor Complex Frequency, Loss of Enteric Neurons, and Intrinsic Primary Afferent Neurons
Understanding which neurons in the ENS are responsible for the frequency of neurogenic contractions along the colon and their direction of propagation has been a major unresolved issue. A major observation of this study was that ablation of calretinin neurons increased CMC frequency. This is opposite to the effects observed in murine experimental models that cause non-specific (cholinergic and non-cholinergic) myenteric neuronal losses in the colon. These include chemotherapeutics (Wafai et al., 2013;McQuade et al., 2016;McQuade et al., 2017;McQuade et al., 2018;Stojanovska et al., 2018;McQuade et al., 2019), colitis (Robinson et al., 2017;Hofma et al., 2018), and disrupted endothelin-3 signalling (Ro et al., 2006;Roberts et al., 2008;Barnes and Spencer, 2015), all of which decrease CMC frequency. Pharmacological blockade of nitric oxide signalling is well known to increase the CMC frequency (Fida et al., 1997;Powell and Bywater, 2001), while the effect of genetic knockout is equivocal (Dickson et al., 2010;Spencer, 2013). Thus, loss of nitric oxide signalling provides a possible explanation for the increases in CMC frequency in the present study. However, this seems unlikely, since although their numbers were halved (Figure 2O), the loss of NOS-IR calretinin neurons as a proportion of all Hu-IR neurons was small and not statistically significant (15 ± 2% vs. 13 ± 1%; P = 0.145; Figure 2H).
The most significantly reduced cell population in CAL-DTR mice were CAL-IR neurons lacking NOS, which is consistent with the neurochemistry of Dogiel type II IPANs (Furness et al., 2004). About 99% of myenteric Dogiel type II neurons (IPANs) express calretinin (Furness et al., 2004), and few express NOS (Hibberd et al., 2018b). IPANs have been suggested to initiate CMCs (Bayguinov et al., 2010;Smith and Koh, 2017), possibly via mucosal 5HT release (Heredia et al., 2013;Smith and Gershon, 2015). Regardless of whether IPANs initiate CMCs, there is direct evidence that IPANs participate in the neuronal firing pattern underlying CMCs (Spencer et al., 2018). There are few studies in which experimental interventions causing neuronal loss or loss of function result in an increase in CMC frequencies. Those that have include P2 × 2 receptor subunit knockout (DeVries et al., 2010) and NaV1.9 knockout (Copel et al., 2013). Interestingly, virtually all Dogiel type II neurons in mouse colon express P2 × 2 receptors, and a subset express NaV1.9 (Copel et al., 2013). Conversely, specific enhancement of IPAN excitability by luminal administration of a neuroactive microbiota has led to decreased intestinal motor complex and colonic propagating contraction frequencies (Wu et al., 2013;West et al., 2017). Thus, the effects on CMC frequency observed with calretinin neuron ablation in the present study may reflect disruption of colonic IPAN function and/or the cells that receive their synaptic outputs. We recently characterised myenteric varicosities arising from IPAN "baskets"-dense clusters of varicosities that represent putative synaptic outputs of IPANs in mouse colon (Smolilo et al., 2019). Most IPAN varicosities (69%) contained calretinin without NOS. Additionally, 35% of IPAN baskets surrounded CAL-IR myenteric nerve cell bodies (Smolilo et al., 2020). Confirmation or rejection of a role in CMC pacing by IPANs awaits more specific pharmacological or neurogenetic targeting of this population.
In a previous study, channelrhodopsin-mediated activation of calretinin neurons evoked premature CMCs and colonic propulsion (Hibberd et al., 2018a). It is interesting then that a reduction of the calretinin neuron population may lead to increases in the frequency of CMCs. We speculate that activation of calretinin neurons is sufficient but not necessary to evoke CMCs. That is, neurons that lack calretinin may also trigger CMCs. In support of this, CALR-IR myenteric neurons represented only 28% of directly mechanosensitive neurons in the mouse distal colon, despite the majority of mechanosensitive neurons being cholinergic (Mazzuoli and Schemann, 2012). This implies numerous other cholinergic neurons capable of directly transducing mechanical stimuli-a major stimulus of CMCs, and is consistent with redundancy of functions in the ENS, which is capable of withstanding major neuronal deficits whilst still showing neurogenic motility patterns (Ro et al., 2006).

Colonic Dimensions and Inflammation
Calretinin neuron losses coincided with a significantly reduced colonic length in the present study. Inflammatory states typically show decreased colonic length (Campaniello et al., 2017;DiCello et al., 2018;Hofma et al., 2018), but indications of inflammation were not present in CAL-DTR colons. Interestingly, a lack of inflammation was also seen using an inducible diphtheria toxin subunit A (DTA) neurogenetic approach to ablate enteric glial cells (Rao et al., 2017), despite inflammation arising in other approaches to ablate glia (Bush et al., 1998). In comparison, they concluded the lack of inflammation following ablation of glial cells in their study was attributable to the DTA method. This is compatible with the finding in the present study that neuronal cell ablation can be achieved using the related DTR approach, without inducing inflammation. Other mouse models that have shown decreased colon length in the absence of overt inflammation include disruption of toll-like receptor 4 signalling (Riehl et al., 2015), and knockout of adenosine A3 receptors (Ren et al., 2011). Whether these mechanisms relate changed colonic length and calretinin neuron losses in the present study remains to be shown.

CONCLUSION
The effects of calretinin ablation in the present study suggest roles in regulating the frequency, force and direction of migration of CMC contractions along the mouse colon. The present study implicates enteric calretinin neurons in the overall regulation of colonic motility and gross morphological structure.

DATA AVAILABILITY STATEMENT
The original contributions presented in the study are included in the article/supplementary material, further inquiries can be directed to the corresponding authors.

ETHICS STATEMENT
The animal study was reviewed and approved by Animal Welfare Committee of Flinders University and Animal Studies 81 Committee at Washington University School of Medicine.