Peripheral Focused Ultrasound Stimulation (pFUS): New Competitor in Pharmaceutical Markets?

Abstract

A new study published in Nature Communications outlines our group’s results using focused ultrasound stimulation within peripheral organs to precisely activate autonomic nerve circuits. The concept is demonstrated by modulating two different (and potentially therapeutic) targets in animal models, a neuroimmune connection in the spleen (that modulates blood cytokine concentrations) and a nutrient sensory pathway within the liver (that modulates metabolism). Connected to this work is a companion Nature Communications publication that utilizes an ultrasound stimulus focused on the spleen to reduce disease severity in a serum-transferred rodent model of inflammatory arthritis. These reports highlight the growing evidence that ultrasound energy (previously shown to enable activation or modulation of central nervous system pathways) may be used to perform peripheral neuromodulation.

In this commentary, we highlight the main findings and discuss their implications for new forms of ultrasound-based therapy. Though challenges remain, a new noninvasive method for precision neuromodulation could solve many of the challenges facing the nascent field of bioelectronic medicine. That is, the use of ultrasound to directly modulate neurophysiological systems therapeutically may provide alternatives to traditional pharmaceuticals. However, to alter the current pharmaceutical paradigm, the field will need to develop a new understanding of how traditional drug concepts (such as dose and pharmacokinetics-pharmacodynamics) relate to the parameters, protocols, and outcomes of this new stimulation technology.

Introduction

Bioelectronic medicine is an emerging field where medical devices modulate nerve circuits that control biological processes.^1,2 Recent clinical trials indicate that these nerve circuits may be targeted and stimulated to treat disease.² However, the development of new therapies relies on gaining a detailed understanding of how the molecular components of the neural pathway are associated with the disease. Also, successful translation to the clinic requires the development of tools to stimulate the pathway in a practical and reliable manner for therapeutic effect.¹

Recent progress in optical and genetic tools has provided a new level of resolution and molecular specificity for mapping and recording from nerve pathways (compared with traditional electrical devices and implants).^1,3 This new level of resolution has increased our understanding of neural reflexes and expanded the repertoire of mapped neural pathways that may be targeted for therapeutic intervention. This list now includes more detailed maps of innervation within peripheral organs, including nerves that modulate inflammatory, metabolic, endocrine, gastrointestinal, and other homeostatic processes. However, despite improvements in molecular specificity, these new tools still have significant scientific, regulatory, and technical hurdles to overcome for clinical translation. Full realization of the promise of bioelectronic medicine may rely on new technologies that stimulate and activate these nerve pathways using practical and clinically applicable techniques and equipment.

Recently, we showed that peripheral focused ultrasound stimulation (pFUS) begins to address this unmet need by enabling targeted neuromodulation directly at innervation points within organs (having specific synaptic connections and physiological function).⁴ This approach leverages the natural hierarchical structure and organization within the nervous system and utilizes focused ultrasound^5,6 to precisely modulate nerve pathways from the level of anatomical collections of nerve endings and synapses (at the millimeter scale within organs). The approach was demonstrated on two different organ systems, including a canonical neuroimmune interface within the spleen (previously described as the cholinergic anti-inflammatory pathway [CAP]^7-11) and an anatomical site containing metabolic sensory neurons¹² within the liver. Two other reports also show that ultrasound stimulation of the spleen alters inflammation in an acute injury¹³ and chronic disease¹¹ model.

Future Clinical Translation and Application

Although preliminary, results from these first experiments with pFUS showed that (in animal models of inflammation) the ultrasound-based approach provided an equivalent reduction in blood cytokine concentrations (i.e., a measure of CAP activation) compared with standard cervical vagal nerve stimulators (i.e., electrical implants).⁴ The two other reports have also shown that alternative methods of ultrasound stimulation of the CAP produce therapeutic outcomes in disease models of inflammatory arthritis and ischemia-reperfusion injury (a mouse model of acute kidney injury).^11,13 Recently, an implant-based approach to CAP stimulation has been translated to a human study, which demonstrated inhibited production of inflammatory cytokines and a positive effect on disease severity scores in rheumatoid arthritis (RA) patients.² Historically, patients with RA (that fail to adequately respond to first-stage treatments such as methotrexate) have then been treated with biologics that inhibit the cytokine tumor necrosis factor (TNF; the same cytokine shown to be modulated using the vagus nerve stimulation [VNS] and pFUS technique^7-10). Although additional data in larger clinical trials will need to be performed to fully assess clinical efficacy, these initial findings show that the molecular target of a major class of pharmaceuticals (i.e., TNF inhibitors or anti-TNFs) may be modulated using a medical device.²

Further advancement in the field will require tools that are able to stimulate and investigate other therapeutically relevant nerve pathways, and these tools will increasingly require the precision to modulate specific pathways without affecting neighboring nerves and tissues.¹ We demonstrate the first study in which pFUS is utilized to stimulate a small suborgan hepatic location containing sensory neurons.⁴ The ultrasound stimulus in this location is shown to modify the sensory input into the major brain centers controlling metabolic homeostasis, and suppress the hyperglycemic effect of endotoxin exposure. Furthermore, the team showed that (unlike standard implant-based VNS¹⁴) the effects of ultrasound stimulation on endotoxin-induced inflammation versus endotoxin-induced hyperglycemia were separable (i.e., hepatic pFUS affected glucose levels and splenic pFUS affected cytokine levels without crossover effect, as is shown with electrical implants).

The development of any new therapy relies on some understanding of the molecular or cellular basis of the disease.¹ Pharmaceutical drugs are often characterized by a high level of molecular specificity (i.e., a chemical or biologic that binds to a specific protein or molecule); however, drugs are most often given systemically and lack anatomical specificity. Often, off-target interactions in other tissues or cells are the root cause of toxicity and side effects. Our initial results⁴ suggest that there may be a new method of modifying the concentrations of important target molecules in a precise manner. This can be achieved by applying targeted pFUS and leveraging the natural organization of the nervous system and the a priori interactions between target molecules and neuro-signaling molecules in the stimulus location. Targeted ultrasound modulation is a novel way of utilizing medical devices, as instruments capable of eliciting spatially distinct, anatomically defined, transient bursts of molecular and cellular activity with suborgan precision.

Devices Cross the Boundary: Defining New Medical Devices Using Drug-Related Methods and Terms

As with any new therapeutic concept, translation to the clinic will hold challenges. Primary among these will be the investigation of a new type of therapeutic “dose.” In our previous report,¹ we demonstrate a dose–response-like relationship between applied ultrasound energy and nerve signaling (i.e., norepinephrine levels) in the spleen. As shown in Figure 1A (and described in more detail in the original report), nerve activation within the spleen and release of norepinephrine triggers a downstream signaling cascade that eventually results in the suppression of TNF release from splenic macrophages. Interestingly, we found that in addition to a dose–response-like analogy (i.e., nerve activation dependence on the applied ultrasound energy level), there also exists a pharmacokinetics–pharmacodynamics (PK-PD)-like analogy between pFUS neuromodulation and drugs. That is, one-time activation of the splenic pathway with ultrasound results in a significant (but decreasing) level of TNF suppression when applied up to 48 h prior to a lipopolysaccharide (LPS) challenge. The persisting effect after the preliminary ultrasound stimulus suggests that the pFUS-induced norepinephrine signaling within the spleen triggers downstream phenotypic changes in the resident immune cells (independent of a preceding inflammatory insult), and this effect lasts long after the original burst of norepinephrine. Our companion paper from Zachs et al.¹¹ also showed a similar response to splenic ultrasound energy in the arthritis model (i.e., a U-shaped response curve to the applied ultrasound pressure), which provides additional evidence of the existence of a therapeutic dosing range of applied energy.

Figure 1.

(A) Schematic of the new form of precision or organ-based neuromodulation outlined in Cotero et al.⁴ In the report, we demonstrate that innervation points of known axonal populations can be targeted for stimulation using focused pulsed ultrasound, including innervation points within the spleen previously shown to harbor an important neuroimmune connection, known as the CAP. The cells outlined in the image are part of this important anti-inflammatory reflex and are described in further detail in our original report. The color-coded key (right) denotes data in panels B and C from different ultrasound treatment versus sham groups in experiments on the LPS-induced inflammation rodent model. Our original publication outlines the methods and details of both the model and ultrasound stimulation parameters in detail. Herein, we expand the original data within the Nature Communications paper¹ to include mRNA expression data from splenic tissue. In addition, we demonstrate data from the model where the LPS exposure was performed after 1 or 2 weeks of daily ultrasound stimulation (with stimulation timing taking place each day at the same time and using the same parameters as shown in the original report). (B) The mRNA expression of TNF in samples from four treatment groups (n = 6) demonstrates a reduction in TNF expression post-LPS injection in Sprague Dawley rats receiving 2 weeks (2w, green) of splenic ultrasound stimulation compared with just 1 week (1w, yellow), or LPS-only (no stimulation, purple) and naïve (gray) controls. (C) The heatmap presents the mRNA expression of 10 genes: TNF (Tnf), interleukin 1 alpha (Il1a), interleukin 1 beta (Il1b), interleukin 6 (Il6), C-C motif chemokine ligand 4 (Ccl4), CCAAT/enhancer binding protein alpha (Cebpa), CCAAT/enhancer binding protein beta (Cebpb), CD14 molecule (Cd14), lipopolysaccharide binding protein (Lbp), and Toll-like receptor 4 (Tlr4). Each column of the heatmap presents the mRNA expression level of the genes for an individual sample. The 24 columns represent the six animals for each of the four treatment groups. The mRNA transcript levels of multiple proinflammatory cytokines were impacted by 2 weeks of ultrasound stimulation versus only 1 week including Tnf, Il1a, Il1b, Il6,¹⁵ and Ccl4 (also known as macrophage inflammatory protein-1β). The proteins of the CEBP family, including those encoded by Cebpa and Cebpb, are important in the development and differentiation of myeloid cells. For example, transcription factor CEBP-beta is important in the differentiation and functioning of TNF-producing macrophages,¹⁶ and the transcription factor CEBP-alpha is involved in macrophage activation.¹⁷ The proteins CD14, LPS binding protein, and Toll-like receptor 4 all have a fundamental role in the detection of LPS and the activation of innate immunity. The samples were harvested from the spleen site as described in the Nature Communications paper.⁴ RNA extraction and RNA sequencing were performed at the Roswell Park Cancer Institute, Genomics Shared Resource, Buffalo, New York.¹⁸ Sequencing mRNA libraries were prepared using the TruSeq Stranded mRNA Sample Preparation Kit (Illumina, San Diego, CA), according to the manufacturer’s instructions. Sequencing was performed on the Illumina NextSeq 500 platform that output 75-base pair pair-ended reads with an average of 50 million reads per sample. Base quality control was checked using Fast QC v0.10.1 from Babraham Bioinformatics.¹⁹ Sequencing reads were mapped to the annotated rat genome version, Rattus norvegicus Rnor 6.0.91, using STAR_2.5.3a aligner.²⁰ Transcript abundance estimates were then generated using RSEM²¹ and output as transcripts per kilobase million (TPM).

Herein, Figure 1B shows an important additional observation in investigating and defining this new pFUS meaning of dose. That is, when measuring the local mRNA expression of TNF in splenic tissue samples (vs the local protein concentration, as reported in our original report), there is a significant difference between the expression in those animals stimulated daily for 1 week versus daily for 2 weeks prior to LPS challenge. In our original report, we show that the splenic and blood TNF protein response to LPS can be nearly eliminated through a single prechallenge pFUS activation of the CAP.⁴ However, this new investigation of the expression level of relevant mRNA within that same tissue after repeated pFUS treatments reveals that there may be important differences between infrequent versus chronic stimulation. That is, for the acute stimulation experiments (i.e., pFUS stimulation for less than 1 week daily) TNF transcript expression is upregulated by LPS challenge to the same extent as the no ultrasound controls. The immediate protein secretion response versus the transcription-level response within the CAP is thought to be mediated by different downstream mechanisms.^22,23 The secretion response is thought to be modulated through the suppression of enzymes associated with TACE^22,23 (the protein in the cell membrane responsible for TNF secretion), and the transcription response is modulated through NF-κB signaling pathways.²³ These results show that although TNF protein secretion is suppressed by acute pFUS stimulation, the transcription-level response remains unchanged. However, chronic stimulation of the pathway daily for longer than 1-week periods resulted in a suppressed TNF transcription-level response compared with the acute stimulation or no stimulation groups. This indicates that exceeding an optimal stimulation threshold may result in compensatory effects on the pathway itself, altering concentrations of key molecules within the pathway, and thus the capability of the anti-inflammatory pathway to respond.

This initial observation opens a wide range of potential future investigations, and druglike comparisons that need to now be investigated. For instance, the potential ability to stimulate and activate acute (i.e., nerve activity and cytokine or hormone secretion) versus long-term (i.e., gene expression and transcript levels) effects within tissue at very precise anatomical locations is unique. It is difficult to systemically administer a drug and yield therapeutically relevant concentrations of that drug only at specific, local anatomical locations. However, we show that precision neuromodulation can be used to rapidly alter local neuro-chemical concentrations and drive therapeutic interactions with local “drug targets.” In addition, it is difficult to chronically administer a drug in a very defined anatomical location over long periods of time. However, we show that precision neuromodulation can be used to alter target molecule activity and/or expression, and that this effect might be tailored based on the frequency of stimulation and activation of local neurons. With this newfound technique, there is now much to be explored. For instance, it is crucial to identify the precision necessary for effective ultrasound administration compared with the natural structure and organization of the organ and innervating nerves. Additionally, it will be important to determine the effect of local stimulation on other surrounding and communicating tissues, and how this might change in healthy versus disease states. Figure 1C identifies some additional findings from the RNA sequencing study described in this commentary and shows that chronic splenic pFUS stimulation impacts more than just local splenic tissue TNF expression. The heatmap shows expression levels of several key cytokines and LPS inflammation-related genes in spleen samples, taken from naïve/unstimulated, LPS/unstimulated, LPS/1-week ultrasound-stimulated, and LPS/2-week ultrasound-stimulated cohorts. Interestingly, there is significant clustering of the naïve and chronic/2-week stimulated cohorts versus the LPS/unstimulated and acute/1-week stimulated cohorts. This further demonstrates the ability of the ultrasound stimulus to be dosed in a manner that changes either the immediate response to an inflammatory event (i.e., TNF secretion alone) or the potentially long-term capability to respond to an inflammatory event (i.e., expression of inflammation/immune-specific genes). Also, compared with the TNF data alone (Fig. 1B), this larger data set shows that the suppression in gene expression due to chronic stimulation is apparent across other molecular components of the innate immune response.

Looking forward, additional data sets using pFUS as an experimental tool will reveal both the extent of nervous system involvement in the chronic disease itself (i.e., how much of the signaling along the CAP is involved in the symptoms and/or complications associated with a disease such as RA) and the ability for this neuromodulation of traditional drug targets to affect diseases in unique ways (i.e., how does pFUS-driven versus anti-TNF biologic drug reduction of circulating TNF affect disease severity in RA models and patients?). Finally, while implants require surgery for placement on a nerve (as in VNS techniques), once implanted a patient can be stimulated or dosed in an automatic and physician-prescribed manner. This may offer an advantage over other forms of therapy in that patient compliance to the therapeutic protocol is built directly into the device (i.e., a stimulation is automatically performed at specific times and frequencies). Patient compliance with a noninvasive stimulation device may depend on the form the final device or product takes, and the ease of use or side effects compared with the patient’s current forms of therapy. Understanding these market and product aspects of the technology will require another set of investigations, in addition to the important scientific and clinical questions posed above. Our group plans to continue to work with the larger clinical and scientific community to safely and ethically perform the investigations required to answer these questions.