Cascade Immune Mechanisms of Protection against Mycobacterium tuberculosis (IMPAc-TB): study protocol for the Household Contact Study in the Western Cape, South Africa

The primary goal of the Cascade Household Contact study is to carry out a comprehensive clinical assessment and collect biological samples from study participants representing a spectrum of TB clinical phenotypes that are defined by 18F-Fluorodeoxyglucose (FDG) positron emission tomography-computed tomography (PET-CT) and immunologic criteria. Classification of the participants into different phenotypic groups will allow results to be compared between groups on different analytical platforms. Participants will also be classified according to evidence of other respiratory infections, including Severe Acute Respiratory Syndrome Coronavirus 2 (SARS-CoV-2) and other common respiratory viruses. Samples from those who have had previous SARS-CoV-2 infection will be analysed for any potential impact on host immune responses to Mtb.

The Cascade Household Contact Study is an observational, case–control, analytical study on human subjects.

Study participants will be enrolled from several areas in Cape Town, Paarl, and the West Coast region in the Western Cape Province, South Africa. These communities carry a high burden of TB disease, up to 880/100,000 population in some areas, and are characterized by poverty, overcrowding and limited health care resources.

The Stellenbosch University Immunology Research Group (SU-IRG) together with the Biomedical Research Institute Clinical Team (BMRI CT) will be responsible for all study participant recruitment and sample collection. The blood and bronchoalveolar lavage (BAL) samples collected will be used for comprehensive analyses of the tissue-specific and systemic immunity that will include cytometry by time-of-flight analyses (CyTOF), RNA-sequencing (RNA-seq), and multiplex immunoassays, epigenetic analysis, and mechanistic studies of Mtb control. Some aliquots of the samples collected through phlebotomy and BAL will be analysed at Stellenbosch University, whereas others will be shipped to partners of Cascade IMPAc-TB project at Oregon Health and Science University (OHSU), SCRI, and the University of Washington (UW) for further analyses (to be described elsewhere).

A Community Advisory Board (CAB) consisting of members of the communities from which participants are recruited meets once a month. The board functions as a forum to advocate for human rights and ethical treatment of research study participants and may contribute to addressing grievances about the research process, give advice on recruitment and retention of study participants, and voice community concerns about study procedures or outcomes. All protocols and informed consent documents are reviewed by the board before implementation. The members of the board also help raise awareness of TB and research in the community.

A total of 236 HIV negative adult study participants with no previous history or serological evidence of SARS-CoV-2 infection (20 subjects for groups 1–8 described in Fig. 1 plus 76 additional participants for follow-up studies) will be recruited and enrolled. TB-index cases will be identified from clinics in and around Cape Town, South Africa with the help of community-based research workers (CRW), and with their consent, their contacts will be approached.

In addition, a total of 80 participants with evidence of previous SARS-CoV-2 infection will be recruited to the study for groups 9–12 (Fig. 1). These participants will be identified through a combination of passive recruitment from amongst those who test positive in other ongoing clinical studies, and active recruitment from the local testing center at Tygerberg Academic Hospital and the National Health Laboratory Services. A previous positive SARS-CoV-2 polymerase chain reaction (PCR) or current positive SARS-CoV-2 serology test will be considered evidence of previous SARS-CoV-2 infection. Testing for SARS-CoV-2 antibodies will be done at screening and repeated within 7 days before bronchoscopy is done in contacts and controls. As vaccination rates increase in the population, it is anticipated that vaccine-induced antibodies will be detected in a growing number of participants and thus the month 3 SARS-CoV-2 antibody tests will be done with both anti-nucleocapsid antibodies (Architect® SARS-CoV-2 IgG, Abbott Laboratories, Sligo, Ireland) and anti-spike protein antibodies (Architect® SARS-CoV-2 IgG II Quant, Abbott Laboratories, Sligo, Ireland). Interpretation of these results combined with history of infection and/or immunization to determine inclusion into the SARS-CoV-2 positive group, will be done as summarized in Table 1. SARS-CoV-2 Real-Time PCR tests on nasopharyngeal swabs will only be done within 7 days before bronchoscopy for infection transmission prevention.

Participants must fulfil all the inclusion and exclusion criteria to be eligible for enrolment (Table 2).

Participants will first be stratified into the following four arms (Fig. 1):

Further classification of contacts and controls within these arms will be based on two IGRA (QuantiFERON-TB Gold Plus® (QFT) assay, Qiagen, Carnegie, Australia) results, performed 3 months apart, as well as the outcomes of PET-CT done within 7 days of the month 3 IGRA test. A combination of TB minus nil 1 and TB minus nil 2 values are used to interpret the QFT test. Should both values be > 0.5 IU/ml, the participant will be classified as IGRA positive. Should both values be < 0.3 IU/mI, the participant will be classified as IGRA negative (< 0.3 IU/ml). If both values are between 0.3 and 0.5 IU/ml or one value is < 0.3 IU/ml and the other between 0.3 and 0.5 IU/ml, the participant will be excluded due to QFN values in the “grey zone”. In cases where one value is equal to or above 0.5 and the other between 0.3 and 0.5, the participant will be included as IGRA positive.

The final classification will further divide the cohort into the following groups (Figs. 1 and 2):

Individuals indicating interest in participating in the study will be pre-screened by CRWs using an online REDCap (Research Electronic Data Capture) [7, 8] based screening tool designed to ascertain suitability for screening. Potentially eligible participants will then be transported to the study site on the Tygerberg Medical Campus of SU. Informed written consent will be obtained before any study procedures are undertaken. All further evaluations will be conducted by the BMRI CT at the Tygerberg Medical Campus and will be scheduled according to the designated arm for the participant.

Demographic data along with participants’ medical, surgical, medication and drug usage history will be collected, and clinical examinations and point-of-care tests documented. All participants will have point of care tests for hemoglobin, random blood glucose and HIV. Females of childbearing potential will also have a urine β-hCG pregnancy test (Homemed, Assure Tech, Hangzhou, China). For participants with a random blood glucose of ≥ 7 mmol/l, a point-of-care HbA1c will be done to exclude Diabetes Mellitus. All participants will receive pre-and post-counselling for HIV tests and positive tests will be confirmed with a second rapid test. All discrepant tests or twice positive tests will be confirmed by a Western Blot Elisa HIV test. Participants excluded due to a new diagnosis of HIV or pregnancy will be referred to the appropriate community healthcare facility for further management.

Sputum, blood, and BAL will be collected from cases, controls, and contacts. This includes those with evidence of sub-clinical infection based on PET-CT. Contacts will be further categorized by their different immunologic response to Mtb antigens over 3 months (indicative of past or present Mtb infection), into those persistently IGRA-negative, persistently IGRA-positive, IGRA-converters (IGRA-negative to IGRA-positive, indicating recent infection), and IGRA-reverters (IGRA-positive to IGRA-negative, suggesting possible clearance). These groups will be compared to one another and controls. In each group, lung immune responses will be linked to blood transcriptional signatures and a comprehensive phenotypic and functional analysis of individual blood immune cells, including Mtb antigen-specific T cells and innate cells.

All participants will be assessed for current or previous SARS-CoV-2 infection, and a group of participants who test positive will be used to assess the potential impact of recent SARS-CoV-2 infection on results.

Blood sampling will take place at the study site according to the schedule for the specific group allocation for the participant summarized in Tables 3, 4 and 5. Following the guidelines set out in Table 6, no more than 140 ml of whole blood will be collected within 3 months. For TB patients, no more than 130 ml of whole blood will be collected within 5 days.

Peripheral blood mononuclear cells (PBMCs) will be isolated and cryopreserved by SUN-IRG for further assays. Proposed assays will be conducted on unfractionated PBMC as well as selected sorted cell populations and analyzed at SU, OHSU, SCRI, UW and other partners. Assessments will include phenotyping by CyTOF or flow cytometry, limited dilution T cell cloning, T-cell receptor (TCR) sequencing, single-cell or bulk RNAseq, multiplex immunoassays and functional granuloma assays to assess differences (e.g., cytokine production, antimicrobial function) between study groups.

The remaining PBMCs will be stored at the BMRI Biorepository Unit (Fig. 3), [9] SU for future assays. Blood (2.5 ml) will be collected in PAXgene® RNA tubes and sent to SCRI for bulk RNA sequencing.

Sputum will be used to detect Mtb by GeneXpert MTB/RIF Ultra® assay, Auramine smear and liquid culture [Mycobacterial Growth Indicator Tubes (MGIT), BACTEC 960, Becton Dickinson] and will be collected by either spontaneous sputum production or sputum induction by hypertonic saline. Isolates from sputum cultures of all cases will be stored at − 80 °C for future analysis including strain typing.

One of the unique features of the Cascade study is the collection of paired peripheral blood and BAL samples from comprehensively phenotyped study participants. This will allow the analysis of immune reactions at the site of infection (lung) in addition to the peripheral immune response in the blood. Bronchoscopies will be performed on all SARS Co-V-2 PCR negative active TB participants and all contacts and controls who do not test positive for SARS-CoV-2 by PCR at month 3 and who had QFT TB antigen minus nil 1 and 2 values either < 0.3 or > 0.5 IU/ml on both the screening and month 3 visits.

Bronchoscopies will be performed in the BMRI Bronchoscopy Suite [10], a research-dedicated bronchoscopy suite located at the Tygerberg Medical Campus (Fig. 4). BAL will be performed under procedural sedation by an experienced team of qualified pulmonologists, medical officers, and nursing staff. The 56 m² suite is fully equipped with two EB-530TX video bronchoscopes (Fujifilm, Tokyo, Japan) with 3.2 mm working channels, an Eluxeo lite Processor (Fujifilm, Tokyo, Japan) with Multi Light Technology, a reprocessing unit (Soluscope, Augbane, France), and a drying cabinet (IDESO, Cape Town, South Africa). The suite is also equipped to monitor and sedate participants during bronchoscopy, with emergency equipment (including defibrillator), and an ultrasound unit (Sonosite, Fujifilm, Tokyo, Japan) with linear and curved transducers. The highest standards of infection prevention for personnel and participants have been implemented in the form of a ventilation system that ensures 12 air changes per hour, UV air sterilizers (Medicare Hospital Equipment, Cape Town, South Africa) and by using 3D-printed Powered Air Purifying Respirators (IDESO, Cape Town, South Africa) for all personnel. Strict procedures for reducing risk for SARS-CoV-2 transmission have been implemented, using the risk assessment and preventative measures detailed in Table 7. A 12-lead electrocardiogram (ECG) will be performed before the procedure on all participants with risk factors for cardiovascular disease.

BAL samples will be obtained from diseased areas of the lung in participants in the active PTB group as determined by chest radiograph, from any FDG-avid areas in contacts, and the right middle lobe in all other participants, unless this lobe is not accessible. Local anaesthesia using lignocaine gel in nostrils, spray in the throat and 2% solution to the vocal cords, trachea, and main bronchi (delivered through the bronchoscope working channel), not exceeding a total lignocaine dose of 4.5 mg/kg, will be used to provide additional analgesia [11]. Nasal or oral intubation will be used depending on the ease of access and participant’s comfort level. The airways will be inspected for any visible abnormalities before the bronchoscope is wedged in the target segment. A maximum of 240 ml of warmed sterile saline will be instilled in aliquots of 60 ml, and gently aspirated. Participants will be monitored for clinical deterioration, blood oxygen saturation, ECG changes, and non-invasive blood pressure (NIBP) before, during, and after the procedure until full recovery and will receive supplemental oxygen throughout the procedure. Light to moderate levels of sedation [level 1 or 2 on the University of Michigan Sedation Scale (UMSS); Table 8] will be attained using either midazolam or propofol combined with fentanyl intravenously.

Before bronchoscopy, sterile saline will be passed through the bronchoscope for the subtraction of background DNA contamination during sequence analysis.

Bronchoscopy with BAL is usually considered safe when performed by experienced operators with appropriate training [12]. Complications can include bronchial irritation with bronchoconstriction (1–2%), pneumothorax (0.1–0.16%) [13] and minor hemoptysis (3%). The risk of severe bleeding is low (< 1%). The rate of complications with fatal consequences is < 1:5000, with only one reported case of death due to sepsis attributed to BAL [14]. The risk of bleeding and pneumothorax is lower in bronchoscopy with BAL compared to procedures such as transbronchial biopsy, which will not be performed in this study [15]. All potential risks associated with the procedure will be clearly explained to the participants during the informed consent process. After bronchoscopy, the participants will be monitored in the recovery room by trained personnel for at least 60 min, until fully awake, stable, and alert. In the case of an adverse event, this period will be extended until the participant is stable. In the case of a minor adverse event (i.e., nosebleeds or coughing) the participants will be allowed to return to their homes. Participants will be provided with a post sedation discharge sheet (Additional file 1). Should any complications arise, transport and treatment at the local hospital will be arranged. The participants will be given the contact number of the study nurse or clinician and will receive a follow-up phone call 72 h and 14 days post-procedure. In the case of severe adverse events, the participants will be admitted to the hospital until resolution.

BAL samples will be transported on ice to the SUN-IRG laboratory where processing will commence within 2 h of collection. End assays will include a range of single-cell and in vitro functional analyses, including single-cell RNA sequencing, Assay for Transposase-Assessable Chromatin (ATAC) sequencing, T-cell cloning, CyTOF phenotyping and 3-dimensional culture assays. BAL from SARS-CoV2 serology positive individuals will be assessed by way of similar end assays, to permit investigation into the effect of this viral respiratory infection on host immune susceptibility to Mtb and vice versa.

Month 3 PET-CT imaging of the contact and control arm participants will be performed at the Nuclear Medicine Research Institute Node for Infection Imaging (NII), SU, South Africa. The NII is a dedicated research PET-CT facility [16] hosted by SU within the Central Analytical Facilities and is situated at Tygerberg Academic Hospital. The unit is the only such facility in Africa accredited as a PET-CT Centre of Excellence by EANM Research Ltd (EARL) [17] and was designed specifically for research in infectious diseases such as TB. The NII is equipped with a Vereos PET-CT scanner (Philips, Amsterdam, The Netherlands) as well as a radio pharmacy for on-site synthesis and labelling of radiotracers. The NII is affiliated with the Division of Nuclear Medicine of SU which has more than a decade of academic PET-CT experience. The principle of ‘as low as reasonably achievable’ (ALARA) for radiation exposure will be followed. The dose of FDG is calculated at 2.8 mBq per kg, in the range of 185 mBq (5 mCi) to 257 mBq (7 mCi) and administered 60 min before imaging. The maximum total radiation from the proposed PET-CT scan is less than 1 Rem/10 mSv. This is lower than what is considered as ‘high’ annual radiation dosage and much lower than the maximal permissible annual research exposure of 5 Rem/year [18]) to 257 mBq (7 mCi) and administered 60 min before imaging. The maximum total radiation from the proposed PET-CT scan is less than 1 Rem/10 mSv. This is lower than what is considered as ‘high’ annual radiation dosage and much lower than the maximal permissible annual research exposure of 5 Rem/year [18].

A light meal will be provided after the PET-CT and bronchoscopy visits for all participants as these visits require participants to be fasting for more than 6 h before the procedure.

Specimens collected will be stored in a fully automated, BMRI Biorepository Unit (Fig. 3) at SU [19]. This unit houses the BiOS Hamilton instrument [9], which has a storage capacity of 4–5 million samples based on tube sizes. Samples are stored at a constant temperature of − 80 °C, with a full audit trail and temperature log for each stored sample. The low energy footprint of the BiOS versus that of conventional freezers will result in a total energy saving of approximately 85%. The facility includes a dedicated laboratory for sample delivery and preparation. Automation will allow samples to be selected without temperature fluctuations that could degrade precious specimens. The system can read 1-D and 2-D barcodes, thereby improving sample tracking. It interfaces with information technology infrastructure and laboratory information management systems for complete automation of sample management. The facility is compliant with ISBER and ISO 20387:2018 and follows all ethical and regulatory requirements.

Participant information will in most cases be collected using standardised study questionnaires on password-protected tablet computers. Information will be captured directly into electronic case report forms on the REDCap platform [8], by study nurses and doctors trained on the study protocol. In some cases, such as documentation generated during the bronchoscopy, information will be captured on standardised paper study data collection forms, which will then be entered into electronic case report forms by a data-capturer.

Electronic case report forms will be programmed to provide warning messages and flags for violations of expected data types for dates, numeric data, and missing data entries. Manual data checks will be performed for essential data entries by a data quality control officer, according to the SU Cascade Household Contact study data management plan standard operating procedure (Additional file 2). The officer will log data queries for study nurses and clinicians to resolve, to ensure that missing or incorrect data are corrected in near real-time.

All participants will be assigned a study identification number and no identifying information will be available to study personnel who do not have direct contact with participants. Data will be stored in a secure, password-protected study database, on servers at the Faculty of Medicine and Health Sciences, SU. The REDCap data on the Hermes server is backed up daily to a secure server in a separate building and back-ups are archived at increasing intervals for 2 years.

A data dictionary was compiled (Additional file 3) to facilitate effective data sharing across various platforms and allow for the description of metadata and syntactic and semantic harmonization. All medical history or new medical findings or conditions diagnosed use ICD-10 (International Classification of Diseases, Tenth Revision) codes and these codes are captured in the eCFR (electronic Case Report Form). The Cascade Data Portal hosted by the UW will be the repository for processed data from this study. Of the 770 variables in the Cascade data dictionary, 232 will be shared with this data portal.

Participants with missing data which prevents classification into one of the study outcomes groups will be excluded from the analysis. We will investigate the demographic characteristics of excluded participants for systematic biases.

Differences between study groups for continuous variables will be assessed with t-tests or generalised linear models for symmetrically distributed variables, and Wilcoxon rank sum tests for asymmetrically distributed variables. Differences between proportions will be assessed with Fisher’s exact tests. Complex relationships between continuous variables will be assessed with clustering techniques or network analysis methods. We will use correction methods for multiple comparisons when required.

For secondary analysis we intend to pool data from different study groups.

Sample sizes were chosen to achieve sufficient power to address primary outcomes. Power was estimated using ANOVA power computation for a range of conditions with varying effect sizes, and varying numbers of groups, and samples per group, to account for variables not relevant to any specific group and for missing data due to measurement failure. Overall, there is greater than 80% power for effect sizes (Cohen’s f) that are in the medium to large range (0.2 < f < 0.4) for continuous-valued measures. Note that Cohen’s f is approximately ½ of Cohen’s d, therefore the power is approximately for any single group mean deviating from the mean of the other groups by at least one within-group standard deviation (SD).

Power to identify differences in RNAseq transcriptomic profiles across groups of size n = 20 exceeds 90% to detect the top 100 genes when fold-change exceeds 3. This calculation is based on conservative assumptions that the minimum average read counts among the 100 prognostic genes in the control group is 5, the maximum dispersion is 0.5, and the ratio of the geometric mean of normalization factors is 1; under these conditions, if the total number of genes tested is 10,000 then power exceeds 90% to detect these prognostic genes if their fold changes are at least 3, using two-sided 5%-level negative binomial exact tests at a false discovery rate (FDR) threshold of 20% [20].