Knudsen, Bruce E.

Abstract Selective activation of smaller diameter over large diameter nerve fibers represents a broad challenge in the electrical modulation of neural activity for clinical applications. For example, off-target activation of motor nerve fibers has been implicated in several failed vagus nerve stimulation (VNS) clinical trials by precluding achieving current amplitudes sufficient to activate the therapeutically targeted small diameter fibers. In this study, we evaluated the use of high-frequency (HF) stimulation to block unwanted activation of large diameter fibers evoked by traditional low-frequency (LF) VNS in a large animal (swine) model. We demonstrate that HF can eliminate specific unwanted LF-evoked components from neural recordings with high selectivity – along with their associated off-target physiological responses - while preserving the activation of smaller diameter fibers driving intended physiological responses. Conversely, block of the smaller diameter fibers was only observed at HF amplitudes much higher than amplitudes commonly used for HF block clinically. We found that in addition to blocking the LF-evoked compound action potential, lower HF amplitudes also drive asynchronous activity which can easily be overlooked. Only at high amplitudes did HF block LF-evoked activation without concomitant asynchronous activation. These data have numerous implications for existing and future LF and HF block neuromodulation therapies.

AbstractBackground and PurposeSpinal cord stimulation (SCS) is approved by the Food and Drug Administration for treating chronic intractable pain in the back, trunk, or limbs through stimulation of the dorsal column. Numerous studies have used swine as an analog of the human spinal cord to better understand SCS and further improve its efficacy. We performed high‐resolution imaging of the porcine spinal cord with intact dura mater using micro‐computed tomography (μCT) to construct detailed 3‐dimensional (3D) visualizations of the spinal cord and characterize the morphology of the dorsal and ventral rootlets.MethodsWe obtained spinal cords from Yorkshire/Landrace crossbred swine (N = 7), stained samples with osmium tetroxide, and performed μCT imaging of the T12‐T15 levels at isotropic voxel resolutions ranging from 3.3 to 50 μm. We measured the anatomical morphology using the 3D volumes and compared our results to measurements previously collected from swine and human spinal cords via microdissection techniques in prior literature.ResultsWhile the porcine thoracic‐lumbar spinal cord is a popular model for SCS, we highlight multiple notable differences compared to previously published T8‐T12 human measurements including rootlet counts (porcine dorsal/ventral: 12.2 ± 2.6, 26.6 ± 3.4; human dorsal/ventral: 5.3 ± 1.3, 4.4 ± 2.4), rootlet angles (porcine ventral‐rostral: 161 ± 1°, ventral‐caudal: 155 ± 6°, dorsal‐rostral: 148 ± 9°, dorsal‐caudal: 142 ± 6°; human ventral‐rostral: 170 ± 3°, ventral‐caudal: 22 ± 10°, dorsal‐rostral: 171 ± 3°, dorsal‐caudal: 15 ± 7°), and the presence and count of dorsal rootlet bundles.ConclusionsDetailed measurements and highlighted differences between human and porcine spinal cords can inform variations in modeling and electrophysiological experiments between the two species. In contrast to other approaches for measuring the spinal cord and rootlet morphology, our method keeps the dura intact, reducing potential artifacts from dissection.

Abstract Renovascular hypertension (RVH) affects up to 7% of individuals over 65 years of age. We have previously demonstrated that induction of oxidative stress is an early event in the development of RVH that leads to development of inflammation and renal atrophy. NLRP3 is one of the most well-known sensors of oxidative stress. However, the role of NLRP3 in the development of RVH and subsequent end-organ damage is not well understood. We therefore sought to test the hypothesis that inhibition of NLRP3 activation protects the heart and the kidneys of mice subjected to RVH. In accordance with our previous observations, wild type (WT) mice subjected to RVH developed 26±6% of renal atrophy within 1 week following cuff placement that progressed to almost 100% by 4 weeks. In contrast, although NLRP3-/- mice (KO) subjected to RVH developed similar degree of atrophy to their age-matched WT in the first week (29 ± 17 vs. 39 ± 20), the atrophy did not progress when evaluated at 4 weeks (29 ± 13 vs. 78 ± 10, p=0.013). Histologic evaluation of renal cortex of KO mice showed no significant interstitial infiltration with CD3+ T cells or F4/80+ macrophages. In addition, KO mice developed no cardiac hypertrophy, no cardiac fibrosis, or aortic lesion despite similar elevation in systolic blood pressure to the WT mice. We propose that NLRP3 activation plays a critical role in the development of renal atrophy in mice subjected to RVH and represent a potential therapeutic target.

Introduction: The estimated prevalence of diabetes mellitus (DM) is 8.4% among adults in the U.S. Cardiovascular disease is a major complication of DM. We have observed that diabetic mice, but not non-diabetic mice, undergoing 2kidney-1clip (2K1C) surgery inducing renovascular hypertension had high mortality attributed to ruptured aortic aneurysm. However, the effect of diabetes on formation of aortic aneurysm remains unclear. Method: In our previous experiment, 163 mice were divided into 4 groups: leptin KO 2K1C (db 2K1C, n=52), leptin KO Sham (db Sham, n=20), wild type 2K1C (dbc 2K1C, n=62), and wild type Sham (dbc Sham, n=29), respectively. We found that 42% (22/52) of db 2K1C died, from ruptured aortic aneurysms examined by gross examination, within 120 days, while others survived 100%. In this study, available H&E aortic sections of survived mice (n=74): db 2K1C (n=30), db Sham (n=15), dbc 2K1C (n=20), and dbc Sham (n=9) were examined 2K1C surgery in a blinded fashion. Aortic lesions defined as necrosis, thrombosis and dropped-out smooth muscle cells, which are signs of early formation of aortic aneurysm, were identified. Inflammation was graded as mild, moderate and severe inflammation. Result: The db 2K1C mice had a significantly greater aortic lesions (16/30, 53%) compared to the db Sham (13/15, 20%); and dbc 2K1C mice (1/20, 5%), p=0.002 and 0.0001, respectively. The number of aortic lesions in the dbc 2K1C group was not statistically different from that of the dbc Sham group (p=0.5). Similarly, the number of aortic lesions in the db Sham group was not statistically different from that of the dbc Sham group (0/9, 0%), p=0.125. Moderate-to-severe inflammation in the tunica media was more prevalence in db mice than the dbc mice. Conclusion: We report for the first time an increased incidence of aortic lesions in db 2K1C compared to dbc and sham groups. This finding suggests that diabetes contributes to aortic aneurysm formation in mice with renal artery stenosis, potentially due to induction of inflammation. Further studies to confirm a mechanism are warranted.

Background: Combined renovascular hypertension (RVH) and type II diabetes (DM II) are a major health problem. Both share common risk factors and contribute to cardiovascular morbidity and mortality. We recently observed a high mortality rate in DM II mice (dbdb) with RVH. Purpose: We sought to determine the interactions between renovascular hypertension and DM II on the cardiovascular phenotype in dbdb mice with RVH. Methods and Materials: RVH was established by placing a polytetrafluoroethylene cuff on the right renal artery of C57BL/KS (wild type, WT) or dbdb mice. We divided the mice into four groups: WT sham, WT type RAS, dbdb sham, dbdb RAS. We examined the blood pressure, plasma angiotensin I, renin, heart weight and myocyte diameter. Histologic parameters including inflammation, necrosis and fibrosis were assessed semiquantitively. Results: Blood pressure, Renin, Plasma angiotensin I: RAS produced a similar elevation in systolic blood pressure (SBP), renin mRNA expression, and plasma angiotensin I content in WT and dbdb mice with RAS. Heart weight: Despite similar elevation in SBP, db/db RAS mice showed a greater increase in heart weight than WT RAS mice (163.7 +/- 6.4 mmHg vs 145.7 +/- 2.3; n=23).. Similar trends were observed on normalization of the heart weight to tibial length. Cardiac myocyte diameter: Cardiac myocyte diameter in WT RAS (n=26) was significantly higher than in WT sham (n=15; 16.4 ± 0.3 vs. 15.0 ± 0.3 μm, p=0.0031). Myocyte diameter was higher in dbdb RAS mice (n=19, 14.76 ± 0.35) than dbdb sham (15.94 ± 0.29 μm, p= 0.0126). However, myocyte diameter was similar in dbdb RAS and WT RAS mice (16.4 ± 0.3 vs. 15.9 ± 0.3 μm, p=0.3). Histology: Sham groups had no inflammation (INF) or necrosis (NEC). Interstitial fibrous tissue was increased in dbdb RAS mice (28% by area) vs WT RAS (19% by area). Dbdb RAS mice had more severe INF (severe in 18%, mild in 82%) than WT RAS (severe in 0%, jild in 12%). Severe NEC was found in 32.1% of dbdb compared to 11.5% of WT RAS. Conclusion: The histologic changes and hypertrophy of the heart superimpose in dbdb RAS compared to WT RAS. This indicates that the risk of cardiovascular morbidity and mortality increases synergisticly in the presence of DM II and RVH.

The objective of this study was to define a role for chemokine generation by parenchymal cells versus circulating inflammatory cells in the initiation of renal damage in the 2K1C model of renovascular hypertension. 129 SV mice were subjected to renal artery stenosis (RAS) by placement of a polytetraethylene cuff on the right renal artery. Systolic blood pressure increased within 2 days of surgery. The cuffed kidney developed tubular atrophy, whereas the contralateral kidney developed compensatory enlargement. Immunohistochemical staining of the cuffed kidney revealed less than 2 CD3+ T cells per 200x field, and rare F4/80+ macrophages or CD20+ B cells in the first 2 weeks following RAS, with progressive increases in macrophages and T cells throughout the 6‐week observation period. Despite the relative paucity of infiltrating inflammatory cells, there was a dramatic increase in chemokine generation within the first week following RAS, as assessed by qPCR analysis of renal cortical homogenates, with 25‐fold induction of MCP‐1, 47‐fold induction of RANTES, and 15‐fold induction of MIP‐2. Chemokine generation continued to increase throughout the 6 weeks, during which circulating inflammatory cells infiltrated the cuffed kidney. Increased expression of MCP‐1 was detected in tubular epithelial cells within 1 week of RAS. We conclude that chemokine generation precedes inflammatory cell influx in RAS.

Despite proven efficacy, ARB therapy reduces glomerular filtration in patients with unilateral renal artery stenosis (RAS). The objective of this study was to determine whether ARB exacerbates renal injury in 2K1C hypertension. RAS was established in mice by placement of a cuff on the right renal artery, which reduces blood flow by 70%. Both Hydralazine (H) and Valsartan (ARB) reduced blood pressure to a similar extent. RAS‐mediated cardiac hypertrophy (heart weight 134% sham (S) p<0.001) was prevented by ARB (102% S, p=ns) but not H (116% S, p=0.03); cardiac fibrosis was prevented in both ARB and H. Although ARB did not exacerbate renal atrophy (91.4 ± 2.1% atrophy vehicle (V) vs 75.3 ± 9.5% ARB, p=ns), H reduced renal atrophy (91.4 ± 2.1% V vs 28.4 ± 13.5% H, p<0.01). The contralateral kidney underwent compensatory hyperplasia (123% S; p<0.001), which was blunted with ARB or H treatment. No significant histopathologic alterations were observed in the contralateral kidneys of any group. RAS induced proteinuria in V mice. Urine protein excretion in ARB RAS mice was reduced to levels observed in S mice; H did not significantly alter urine protein excretion. Severe renal atrophy was observed in the cuffed kidneys of V mice, which was not exacerbated by ARB. However, ARB reduced proteinuria to baseline levels, indicating that ARB may protect overall renal function in unilateral RAS. These studies were supported by NIH grant P01 HL85307.

Recent studies have implicated CCR2 ‐ CCL2 signaling in macrophage influx and development of chronic renal disease in human and experimental renal artery stenosis (RAS). We sought to test the hypothesis that development of chronic renal disease would be abrogated in CCR2 knockout (KO) mice. We established RAS in CCR2 KO and WT mice (n=10) through placement of a cuff on the right renal artery We have shown that the cuff reproducibly reduces blood flow by 65% to the kidney. Sham mice underwent manipulation of the right renal artery without cuff placement. Both CCR2 KO and WT mice developed hypertension, which persisted throughout the 4 week study. There was no difference in renal atrophy between CCR2 KO (53+//‐14.6% cortical surface area) and WT (61.2%+/‐14%) mice. Furthermore, there were no significant differences in F4/80+ macrophage infiltration (12.78+/‐4.9% surface area for CCR2 KO mice and 13.13+/‐ 3.1% for WT mice) or CD3+ T cell infiltration. Infiltrating macrophages were predominantly iNOS+ M1 phenotype; few arginase+ cells were observed in any group. Urine protein excretion was similar in CCR2 KO and WT mice. CCL was induced to a similar extent in CCR2 KO and WT mice. Furthermore, no differences in CCL5, CCL7, CCR5, CCR1, or iNOS expression was observed between CCR2 KO and WT mice. We conclude that macrophage infiltration and chronic renal damage occurs in mice with deletion of CCR2. Future studies will identify signaling pathway(s) responsible for macrophage influx and chronic renal damage in this model. Grant Funding Source : AI‐100911

Cervical vagal nerve stimulation (VNS) is an FDA approved treatment for epilepsy and depression and is currently being evaluated for treatment of numerous other disorders including pain, anxiety, and inflammation. VNS also received European market approval for heart failure, bronchoconstriction, and diabetes; however, approval for these indications was based on studies in animals, such as dogs and rats, whose vagal nerve size may not have been representative of humans. When subjected to double‐blinded, randomized, sham controlled studies for subsequent U.S. approval, many of these CE‐marked VNS indications failed to meet their primary efficacy endpoints. More than 100,000 patients have undergone VNS, but despite the efficacy of vagal nerve stimulation in a subset of patients, many fail to reach stimulation levels that result in alleviation of symptoms without triggering unwanted side effects. Post hoc studies of the patient population suggested that VNS may not engage the intended fiber pathways in humans, despite engagement in preclinical studies in mouse, rat, or dog. Stimulation parameters were limited in these studies as higher parameters caused side effects such as cough, voice alteration, and dyspnea. Discrepancies between stimulation amplitude needed for intended effect versus unwanted side effect could be directly related to the relative size of the vagus nerve across animal models, variations in fascicle organization, or variations in branching patterns. To address translating improvements in VNS from the lab to the clinic, we evaluated the swine as a potential model for approximating the clinical environment. The size and fascicular organization of the swine vagus nerve is more representative of the human nerve than previously used models. We used a combination of histology, microdissection, and ultrasound to establish a distinct bimodal organization of groups of fascicles stemming from one large fascicle at the inferior (nodose) ganglion, an organization we have termed ‘vagotopy’. We demonstrated that with the use of ultrasound we can visualize the fascicular organization in vivo , an approach that could be used clinically to target stimulation of either sensory fibers or parasympathetic efferents, depending on therapeutic intent. These advances in a clinically relevant model not only inform electrode placement for target engagement, but provide an approach for reduction of side effects. Support or Funding Information Acknowledgements Funded by TL1TR002380 andThe Defense Advanced Research Projects Agency (DARPA) Biological Technologies Office (BTO) Targeted Neuroplasticity Training Program under the auspices of Doug Weber and Tristan McClure‐Begley through the Space and Naval Warfare Systems Command (SPAWAR) Systems Center with (SSC) Pacific grants no. N66001‐17‐2‐4010, and the NIH SPARC Program Award OT2 OD025340.

Vagus nerve stimulation (VNS) is FDA‐approved and CE‐marked for a diverse set of conditions including epilepsy, depression, heart failure, and others. Despite widespread adoption and perceived efficacy, a post‐hoc analysis of patients treated for heart failure using VNS revealed that only 12% of patients showed stimulation evoked heart rate changes one‐year post‐implant. This finding is not unexpected given that A‐type nerve fibers have the lowest activation thresholds, and they innervate neck muscles, causing contractions that patients will not tolerate. Therefore, muscle responses at lower amplitudes may have prevented application of higher amplitudes required to achieve cardiac effects in clinical trials of VNS for heart failure. Understanding therapy‐limiting side effects to VNS and finding ways to mitigate them are critical for optimizing VNS for clinical therapies. We developed a swine model of VNS which closely approximates the size and fascicular organization of the cervical vagus nerve observed in humans. The clinical LivaNova stimulation electrode was used as advised by a practicing neurosurgeon. We hypothesized that neck muscle responses are mediated by somatic branches of the vagus nerve as opposed to direct electrical activation of the muscle fibers. Further, we hypothesized that stimulation evoked responses depend on the location of the stimulation electrode on the vagus nerve. We measured compound nerve action potentials using intrafascicular electrodes placed within the vagus to identify activation of different nerve fiber types by conduction velocity. The downstream effects of nerve fiber activation, i.e. changes in heart rate and neck muscle responses (EMG), were measured using an arterial catheter and intramuscular electrodes, respectively. We then used a neuromuscular junction blocking agent, as well as systematic transection of somatic vagus nerve branches, to specifically identify the pathways causing neck muscle contractions. Two pathways contributed to VNS‐evoked neck muscle responses: activation of nerve fibers in the recurrent laryngeal branch of the cervical vagus nerve at low amplitudes and in the superior laryngeal branch at higher amplitudes. Lesser activation of the superior laryngeal branch could be achieved by moving the stimulation electrode caudally down the length of the vagus. This work represents a step towards mitigating intolerable side effects of VNS, thereby allowing maximization of intended target engagement and therapeutic benefit. Support or Funding Information The Defense Advanced Research Projects Agency (DARPA) Biological Technologies Office (BTO) Targeted Neuroplasticity Training Program under the auspices of Doug Weber and Tristan McClure‐Begley through the Space and Naval Warfare Systems Command (SPAWAR) Systems Center with (SSC) Pacific grants no. N66001‐17‐2‐4010, and the NIH SPARC Program Award OT2 OD025340.