Peripheral TMD Pain Mechanisms and the Effect by Botulinum Toxin A

Patient recruitment:

The patients will be recruited among those referred to the Specialist Clinic for Orofacial Pain and Jaw Function at the University Dental Clinic, Karolinska Institutet, Huddinge, Sweden, or via advertisement.

Randomization:

Patients will be randomly assigned to treatment by a random generator (www.randomization.com). For each participant the treatment will be written on a note and placed in a sealed envelope by a researcher not involved in any other part of data collection. A second person not involved in patient examination will open the envelope, prepare the syringe with the randomized solution, and place it in the examination room before the examiner and patient enters it. Botulinum toxin and saline have identical appearance so the participant and investigator will be masked to treatment assignment.

Procedure:

The participants complete an extended and slightly modified version of the Swedish Axis II questionnaire included in the Diagnostic Criteria for TMD (DC/TMD) before the first visit. The Axis II questionnaire contains sociodemographic question, questions regarding TMD symptoms and headache (presence, duration, and frequency) used for diagnostic purposes (Symptom Questionnaire), and validated instruments to assess physical and emotional function. A clinical examination is then performed according to the DC/TMD Axis I to ensure that the participant fulfils the enrollment criteria. If so, consent is signed. Then quantitative sensory testing (QST; thermal thresholds, pin-prick pain, temporal summation pain, pressure pain threshold, and conditioned pain modulation) is done and biopsies are obtained from a painful (MM) and a non-painful site (AT) for later analysis. At least one week hereafter treatment are given.

Axis II questionnaire:

For physical function the following instruments are used: PROMIS physical function short form (6b) to assess general physical function, the Oral Behavior Checklist (OBC) to assess presence and frequency of oral parafunctions, and the Jaw Functional Limitation Scale (JFLS).

To assess emotional function these instruments are included: The patient Health Questionnaire (PHQ-8) to screen for depression, the Generalized Anxiety Disorder (GAD-7), the Pain Catastrophizing Scale (PCS), the Perceived Stress Scale (PSS-10), the PROMIS sleep disturbance short form (6a) questionnaire, and the Oral Health Impact Profile (OHIP-5) to assess oral health-related quality of life.

Biopsies:

Microbiopsies will be obtained from one of the MA and the AT on the same side (a non-treated pain-free internal control). The microbiopsies are taken through the skin overlaying the most prominent part of the muscle under skin surface anesthesia. A disposable biopsy instrument with a penetration depth of 10 mm and a diameter of 18 gauche (G) (MM) or 16G (AT) are used. The biopsy instrument will be guided with a coaxial needle, that will be inserted to a depth of 10 mm. Maximum four microbiopsies are taken from each muscle to ensure that sufficient muscle tissue is obtained. The coaxial needle will be inserted along the long axis of the muscles until the fascia is penetrated and the biopsy instrument is then inserted through the coaxial needle and a piece of the muscle with a size of approximately 0.12 cm * 1.1 cm is collected. The muscle section will be removed from the biopsy instrument using a sterile probe and immediately put in a cryotube which is snap frozen in liquid nitrogen. After removal of the muscle section the biopsy instrument will be rinsed with isotonic saline. This procedure will be repeated three times (if possible). Each time the biopsy instrument will be rotated 45° so that the microbiopsy would be taken from a new portion of the muscle. The biopsies will be stored (-80º) in Karolinska Institutet's dental biobank until analysis.

Treatment:

Patients in the active treatment group receive 100 IU of botulinum toxin dissolved in 1 ml of isotonic saline. This dose is far below the lethal dose of 40 IU/kg body weight and is commonly used for treatment of TMD. In each MM 30 IU is injected divided between three different points (origin, mid-portion, and attachment) and in each temporalis muscle 20 IU (anterior and mid-inferior portion) after careful aspiration. Thus, in each point 10 IU of solution is injected. The solution is injected with EMG guidance using an audio-amplified device (Allergan Norden AB) to ensure intramuscular administration. A 1-mL syringe with a 19 mm needle (diameter 0.4 mm) will be used. Patients on the control treatment group receive intramuscular injection with isotonic saline (0.9%) in the same manner.

Analyses of biopsies:

The biopsies will be analyzed with bulk ribonucleic acid sequencing (RNA-seq) and single-cell seq including spatial transcriptomics at University of Texas Health Science Center in San Antonio (UTHSCSA). Flow cytometry (FC) analyses will be done at the ANA Futura lab in Huddinge, Sweden. IHC will be performed at the Live Cell Imaging Facility at Karolinska Institutet, Huddinge, Sweden, and proteomics at SciLife lab in Solna, Sweden. Data of the non-painful site (AT) will be subtracted from data of the painful site (MM) for every patient. Then, these data can be clustered.

For bulk RNA seq, differentially expressed genes (DEGs) will be selected and then paired analysis will be run within empirical analysis of DEGs in R (edgeR) software to select significantly different (p\<0.05) DEGs. This is an established and well standardized approach. Selection criteria for further analysis are fold charge (FC)>1.5 and Reads Per Kilobase of transcript per Million mapped reads (RPKM)>10. Finally, biological processes will be defined by www.pantherdb.org on base of DEGs. Subtraction of internal controls from experimental data will also be done, followed by clustering of patients according to transcriptional profiles using K-mean principial component analysis algorithm. This last analysis is critical to build association of TMDM patient phenotypes with gene plasticity in MM and will be done with artificial intelligence (AI) and data learning (DL) algorithms by the Bioinformatics and Data Science Core (BDSC) at UTHSCSA. ScRNA-seq will be analyzed by SEURAT, which is by far more precise than standard clustering software from 10x Genomics.

The IHC analysis will focus on sensory neurite plasticity using markers for subsets of sensory neurons and use Light Sheet microscopy to measure neurite length and branching for whole biopsies. The following markers will primarily be used: PGP9.5, NFH, MrgprD, HTR3A, Nav 1.8, TrkA, TrkB, TrkC, MrgprA3, SP, CGRP, TRPV1, parvalbumin, calbindin, somatostatin, and PENK, but other markers can be considered as well. We also plan to use the Glyclick method which generates stable and homogenous antibody conjugates for IgG from several species and subclasses. FC will be used to profile immune and vasculature cells.

Proteomics will be done with two methods. First two-dimensional gel electrophoresis followed by identification with liquid chromatography-tandem mass spectrometry, followed by protein network analyses. Therafter affinity proteomics using Olink platforms will be done, which enables analysis of several hundreds of proteins in one analysis.

When all the analyses are done and the results published any remaining material will be discarded.

When the analyses at UTHSC are done, and the results published any remaining material will be discarded. If any material is remaining from samples in Sweden, they will be kept in the dental biobank at KI for other analyses.

Statistical analyses of primary outcomes (biological markers):

Pilot experiments show that 20 patients in each group are needed to attain >80% power for each test detecting at least 2-fold significant difference (FC) in gene expression with an estimated standard deviation of 0.6 (considering high human-to-human variation; this includes samples after botulinum toxin treatment. Power analysis show that we need 10 paired samples per group for scRNA seq, IHC and FC. To obtain sufficient power data from 40 patients will be included in proteomic analyses, which will give a power of 0.856 with an expected difference from baseline at the 1-mo follow-up of 20% (SD 30%) with alpha = 0.050 and beta = 0.20 (paired t-test).

Statistical analyses will be done with 1-, 2- and 3-way and repeated measures ANOVA with Bonferroni's post-hoc test. For scRNA seq, currently accepted analysis for is that paired samples belonging to the same participant group combined and multiple t-test-like analysis for each cell cluster run against another group. Prism v8 software (GraphPad Software LLT, Boston, MA, USA) will be used with p \< 0.05 as significance level.

Statistical analyses of secondary outcomes (clinical data):

Even if the sample size is not calculated based on clinical data, power analysis for pain intensity based on a previous study show that 20 patients per arm would yield a power of 0.977 given a group difference of 3.14 (SD 2.45) and a significance level of p \< 0.05. Given the invasive nature of the project (repeated microbiopsies) a relatively high drop-out rate is expected why 5 additional participants will be recruited per group (MYA, MFP, healthy controls).

For statistical univariate analyses the SigmaPlot v 14 software will be used (Systat Software Inc, Chicago, IL, USA). Mean (SD) or median (interquartile range, IQR), depending on normality, will be used to describe the data. To compare groups in Project A, ANOVA or Kruskall-Wallis' test will be used. Repeated measures ANOVA with Dunnet's posthoc test will be used the analyze differences in treatment effect with Group and time as factors in Project B. Data that are skewed or ordinal (most Axis II data) will be analyzed with Friedman ANOVA for each group separately with Dunn's test as posthoc test for time differences. The Mann-Whitney U-test will be used to analyze differences between groups at different time points, using Bonferroni correction for multiple testing. Differences are considered as significant when p \< 0.05.

Further, multivariate statistics will be used to identify correlations between significantly changed biomarkers and other variables using SIMCA v.17.0 (Sartorius Stedim Biotech, Umeå, Sweden). Principal component analysis (PCA) will be used to identify moderate or strong outliers. Orthogonal partial least squares discriminant analysis (OPLS-DA) will then be used to regress group membership using the biomarkers as regressors. For correlations between biomarkers and clinical variables OPLS modeling will be performed. The variable in-fluence on projection (VIP) value indicates the relevance of each X-variable pooled over all dimensions and the Y-variables indicate the group of variables that best explain Y. VIP > 1.0 is considered significant. R2 describes the goodness of fit, i.e., the fraction of sum of squares of all the variables explained by a principal component whereas Q2 describes the goodness of prediction, i.e., the fraction of the total variation of the variables that can be predicted by a principal component using cross validation methods. R2 should not be considerably higher than Q2, if substantially higher (>0.3) the robustness of the model is poor. To validate the model, an obtained CV-ANOVA p-value will be used. The OPLS-DA model is considered of significant importance if the CV-ANOVA has a p-value \< 0.05.