Peripheral Focused Ultrasound Neuromodulation (pFUS)

Every organ in the body is innervated by nerves that either send signals to the organ (and effect that organs function) or sense the current state of the organ (and communicate back to the central nervous system). (Chakravarthy et al., 2016; Patil and Thakor, 2016; Cogan, 2008; Shemesh et al., 2017; Deisseroth et al., 2015; Olofsson and Tracey, 2017; Birmingham et al., 2014). However, until recently our understanding of those interactions was limited, primarily due to the tools available to accurately stimulate and record from neurons communicating with defined anatomical and functional organ locations (Chakravarthy et al., 2016; Cogan, 2008; Birmingham et al., 2014; Ng et al., 2016). Recently, new optical and genetic tools have provided new ways with which to communicate with and experiment on specific nerve pathways (versus use of traditional electronic implant-based devices) (Shemesh et al., 2017; Deisseroth et al., 2015). With these new tools has come new understanding of the specific roles that nerves play in both the homeostatic and pathological function of the tissues they innervate.

Bioelectronic medicine (BEM) is an emerging field which leverages medical devices to modulate nerve circuits that are found to control important physiological processes (Olofsson and Tracey, 2017; Birmingham et al., 2014). Clinical trials now indicate that these nerve circuits may be targeted and stimulated with therapeutic effect. These include a study in which implant-based vagus nerve stimulation (VNS) demonstrated inhibited production of inflammatory cytokines and a positive effect on rheumatoid arthritis (RA) severity scores (Koopman et al., 2016). Although performed on a limited number of patients, the study demonstrated both response to the electrical stimulation (via DAS28CRP scores) and worsening of disease severity upon withdrawal of the BEM treatment. Additional data in larger studies will be required to assess full clinical efficacy. These initial findings show for the first time that the molecular target of a major class of pharmaceuticals (i.e. TNF inhibitors or anti-TNFs) may be modulated using a medical device.

However, despite recent successes, these BEM tools still have significant scientific, regulatory, and market hurdles to overcome. And, full realization of the promise of bioelectronic medicine may require technologies that activate these nerve pathways using less invasive techniques, while providing an increased level of precision (compared to implants that stimulate many neurons within large nerve bundles). Previously, we demonstrated the first study in which peripheral focused ultrasound neuromodulation (pFUS) is utilized to stimulate sub-organ locations containing specific targeted neurons (Cotero et al., 2019a). Results were shown in two different organ systems, including a splenic target (neurons associated with the cholinergic anti-inflammatory pathway (Tracey, 2009; Pavlov and Tracey, 2012; Tracey, 2016; Wang et al., 2002) and a hepatic target (neurons associated with metabolic sensing and control (O’Hare and Zsombok, 2015; Roh et al., 2016; Pocai et al., 2005; Yi et al., 2010; Miki et al., 2001; Stauss et al., 2015; Karnani and Burdakov, 2011; Routh et al., 2014; Marty et al., 2007). Herein, we expand this concept by 1) confirming direct pFUS nerve activation by applying ultrasound stimuli to an in vitro three-dimensional (3D) nerve culture (Griffin et al., 2015; Tay et al., 2018) (Fig. 1A), 2) demonstrating pFUS activation of neural pathways at multiple strategic peripheral target locations (in addition to the end-organ targets shown previously; Fig. 2(A), and 3) demonstrating pFUS capability to affect a specific pathological state (i.e. LPS-induced hyperglycemia) using multiple associated end-organ stimulation targets (i.e. liver, pancreas, and intestines; Fig. 3A).

In our previous report, we demonstrated that sub-organ stimulation of a specific site in the liver (known to harbor glucose and metabolite sensory cells (O’Hare and Zsombok, 2015; Roh et al., 2016; Pocai et al., 2005) modulates systemic metabolism through a hypothalamic pathway (Cotero et al., 2019a; Routh et al., 2014), while stimulation of the spleen (known to harbor neurons that regulate cytokine concentrations in the blood (Tracey, 2009, 2016; KJ, 2002) modulates inflammation through the cholinergic anti-inflammatory pathway (Cotero et al., 2019a). We demonstrated that these two pathways were stimulated simultaneously when using standard implanted vagus nerve stimulator (VNS) technology, as neurons from both end-organs are stimulated by the implant. However, pFUS enabled separate stimulation of each pathway (i.e. modulation of cytokines/inflammation at the splenic pFUS site without altering blood glucose/metabolism, and modulation of blood glucose at the liver pFUS site without altering cytokines/inflammation status) (Cotero et al., 2019a). Significant data was obtained suggesting pFUS was acting on these pathways by stimulating neurons within the ultrasound field (including ultrasound-induced neurotransmitter modulation in the spleen, elimination of the splenic effect using reserpine denervation, ultrasound-dependent hypothalamic cFos expression after liver pFUS, and verification of ultrasound-activated afferent nerve pathways using DfMRI) (Cotero et al., 2019a). In addition, two other groups have reported evidence of ultrasound-induced activation of the splenic pathway (Zachs, 2019; Gigliotti et al., 2013). However, despite this evidence in pre-clinical models, most attempts to directly record action potentials and ultrasound activation in stimulated peripheral nerves have failed in vitro (Wright et al., 2015; Colucci et al., 2009; Juan et al., 2014).

We hypothesized that past challenges in achieving direct measurements of ultrasound-induced peripheral nerve activation can be attributed to the stimulus target. In the recent reports of in vivo ultrasound induced peripheral neuromodulation, (Cotero et al., 2019a; Zachs, 2019; Gigliotti et al., 2013) stimulation targets were located directly in end-organs (i.e. neurons, axons, and end-axon terminals within organs). However, in past reports that failed to achieve direct ultrasound-mediated nerve activation, the stimulus target was large myelinated or unmyelinated nerve bundles (Wright et al., 2015; Colucci et al., 2009; Juan et al., 2014; Wright et al., 2017). Further validating this hypothesis are previous ultrasound stimulation studies on central nervous system (i.e. brain tissue) targets that demonstrated successful activation of CNS neurons through an US-induced effect on SNARE-mediated synaptic vesicle exocytosis and synaptic transmission (Tyler et al., 2008; Tufail et al., 2011). In the PNS these molecular components (i.e. cell soma, SNARE proteins and synaptic connections) are more prevalent in end-organ sites or ganglia (compared to inter-organ nerve/axonal bundles).

Experiments investigating CNS nerve activation often utilize brain-slice experimental preparations, in which excised sections of brain tissue are temporarily maintained at an artificial interface to fluorescence microscopes and electrodes for analysis of nerve activity (Tyler et al., 2008). However, such experimental in vitro systems for PNS tissue are less common, as peripheral nerves span long lengths in the body and interface with an array of complex tissue types. Typically, PNS experiments have been limited to studying excised axon bundles that have been cut and separated from their soma and synaptic interfaces (Wright et al., 2015; Colucci et al., 2009; Wright et al., 2017). Herein, we first utilize a novel 3D culture system (Griffin et al., 2015; Tay et al., 2018) (Fig. 1A; details in materials and methods) to provide a more optimal in vitro experimental platform for peripheral nerve investigation compared to traditional excised nerve bundle preparations. We observed formation of complex neural networks within the culture platform and utilized the novel system to couple ultrasound transducers and an observational fluorescence microscope to the neurons. Use of calcium indicator dyes enabled direct observation of neuron activity. Using the system, we verified that the ultrasound pressures used previously for in vivo neuromodulation at end-organs (Cotero et al., 2019a; Zachs, 2019) results in direct nerve activation in the culture system.

We then expanded in vivo neuromodulation experiments compared to those previously reported (Cotero et al., 2019a) by stimulating alternative sites known to contain peripheral nerve soma and synapses, including sensory ganglion (i.e. inferior ganglion of the vagus nerve or no dose ganglion) and peripheral ganglion of mixed (sensory and efferent) innervation (i.e. sacral ganglion). We demonstrate activation of the cholinergic anti-inflammatory pathway (Tracey, 2009, 2016; KJ, 2002) (i.e. modulation in LPS-induced circulating cytokine concentrations (Cotero et al., 2019a) after ultrasound stimulation at each site (i.e. end-organ/spleen, no dose, and sacral ganglia) and show that both the magnitude of cytokine reduction and presence of other off-target effects (i.e. simultaneous changes in blood glucose levels) are stimulation site dependent. Lastly, we show intervention in a specific pathological state (i.e. reduction of LPS-induced hyperglycemia) through stimulation of multiple strategic anatomical locations associated with metabolic control. These exploratory stimulation sites included the hepatic site containing peripheral glucose sensors, (Cotero et al., 2019a; O’Hare and Zsombok, 2015; Roh et al., 2016; Pocai et al., 2005; Routh et al., 2014), the pancreas associated with insulin secreting beta cells (Suarez Castellanos et al., 2017), and an intestinal site containing incretin secreting enteroendocrine cells (Sandoval and Sisley, 2015). We show that ultrasound induced attenuation of hyperglycemia is achieved by stimulation at each of the anatomical targets. However, the magnitude of blood glucose reduction is stimulation site dependent, and the effect is driven by different molecular mechanisms at each site. Together this data demonstrates that ultrasound stimuli are capable of direct peripheral nerve (and other neuroendocrine cell) activation, and that as a new tool pFUS is capable of efficient assessment of the effectiveness of alternative therapeutic BEM stimulation sites.