Noninvasive ultrasound stimulation of the spleen to treat inflammatory arthritis

Abstract

Targeted noninvasive control of the nervous system and end-organs may enable safer and more effective treatment of multiple diseases compared to invasive devices or systemic medications. One target is the cholinergic anti-inflammatory pathway that consists of the vagus nerve to spleen circuit, which has been stimulated with implantable devices to improve autoimmune conditions such as rheumatoid arthritis. Here we report that daily noninvasive ultrasound (US) stimulation targeting the spleen significantly reduces disease severity in a mouse model of inflammatory arthritis. Improvements are observed only with specific parameters, in which US can provide both protective and therapeutic effects. Single cell RNA sequencing of splenocytes and experiments in genetically-immunodeficient mice reveal the importance of both T and B cell populations in the anti-inflammatory pathway. These findings demonstrate the potential for US stimulation of the spleen to treat inflammatory diseases.

Conflict of interest statement

The authors declare no competing interests.

Figures

Fig. 1

Modulation of the cholinergic anti-inflammatory pathway through the vagus nerve, splenic nerve and spleen. a Electrical stimulation of the vagus nerve or US stimulation of the spleen is thought to modulate the neural communication with T Cells and Macrophages, blocking the production of inflammatory cytokines and promoting an anti-inflammatory state. b Timeline of a typical experiment performed in the presented study in which animals were injected with 300 μl of K/BxN serum on day 0 and treated with focused US that targeted the spleen on days -1 through 6

Fig. 2

US treatment modulates arthritis severity. Results from 7-day arthritis experiments are presented. a–d Pooled data from three experiments (total n = 32 mice), using identical experimental conditions (1 MHz US at 350 kPa, 1 s on/5 s off bursts for 2 min per day; shallow US-focusing cone to target the spleen, see Supplementary Fig. 2a) are shown. On the final day of the experiment, change in ankle thickness (a) and clinical score (b) of US treated animals are significantly reduced compared to sham-US controls (p = 0.0017 and p = 0.0060, respectively, using a Mann–Whitney Test). For comparison, normalized data from the same set of experiments are shown in c, d. A normalized value of zero indicates that the animal’s arthritis outcome was similar to the average untreated (sham-US) animal, a positive value indicates clinical worsening, and a negative value indicates clinical improvement. e, f Dose-response curves of US pressure reveal an optimal US amplitude between 333–350 kPa. Each point represents the normalized mean change of US treatment compared to sham at day 7. 1 MHz US stimulation with 1 s on/5 s off bursts for 2 min per day was applied in each experiment. Circles indicate that the shallow US-focusing cone was used, and triangles indicate that the deep US-focusing cone was used. Results are pooled from 108 mice across 9 experiments (see Mice section of Methods). g Additional US stimulation parameters and body locations were also tested, revealing consistent and effective therapeutic effects when specifically using 350 kPa, 1 s on/5 s off bursts with the shallow US-focusing cone targeting the spleen. One hundred and three mice were used over 9 experiments. All conditions consisted of 2 min of US stimulation per day, except for the right leg conditions that consisted of 4 min per day. Contra-spleen refers to right abdominal stimulation contralateral to the spleen. Tcra KO and muMt- were spleen stimulation experiments performed in T cell and B cell knockout mice, respectively. Mean and SEM are shown for all figures. Double asterisks denote p < 0.01. kPa kilopascal

Fig. 3

US treatment is effective before or after arthritis onset and depends on duration of stimulation period (page 18). All experiments in this figure used the 1 MHz US at 350 kPa, 1 s on/5 s off bursts in animals with arthritic serum transfer at day 0. a, b Full 7-day time course pooling data from the three experiments shown previously (Fig. 2a, b), which used identical experimental conditions for 2 min per day with US administered daily from day -1 through 6 (total n = 32). c, d Ultrasound duration causes a dose-dependent effect on arthritis. 6, 12, and 20 minute US stimulation was tested and is shown with the 2 min US and sham results from a, b for comparison. Figures show the 7-day progression of the disease and mice were stimulated with US from day -1 through day 6. For the 6, 12, and 20 min experiments, there were 6 animals per cohort (total n = 18) and for the 2 minute US and control data there were 16 animals per cohort (total n = 32). e, f US treatment is also effective at reducing symptoms after onset of arthritis. Treatment was not initiated until day 3 after arthritis had manifested. US was administered for 20 min per day. Data was pooled from two experiments with a total n = 23. g, h Normalized 7th day values of ankle thickness (g) and clinical score (h) for the different durations of ultrasound treatment shown in 3a–f. Error bars are SEM. 6, 12, and 20 minute values were normalized to the pooled sham control data from a, b. The bar “20 min US initiated on day 3” shows the improvement in ankle thickness and clinical score compared to the first day of treatment on day 3. In a–f, whisker plots are presented for better visualization of the data over time in which the endpoints of the vertical lines span the minimum to maximum values, the midline of each box is the median value and each box extends from the 25th to 75th percentile. In g, h, mean and SEM are plotted for each condition. kPa kilopascal

Fig. 4

Splenic T and B cells demonstrate induction of genes following US treatment. The dot plot displays the statistically significantly differentially expressed genes (DEGs) in T cells (top list) and B cells (bottom list) from mice given arthritogenic serum comparing US versus sham-US stimulation. The size of each circle represents the percent of cells within each cell type (T cells = T, B cells = B) which express the gene listed, and the color of each circle represents average scaled expression. Gray bars denote mouse treatment groups; genes that are in bold are statistically significantly DEG in both T cells and B cells, and genes found to also be DEG in non-arthritic mice with US stimulation (see Supplementary Fig. 8) are denoted with a superscript x. Adjusted p-values for each gene in either T or B cells (top or bottom gene lists) are shown in shades of green. All DEGs listed in this Figure range in statistical significance from p = 2.39e-02 to p = 3.95e-230 using a Wilcox rank sum test with Bonferroni correction