Methods, Quality Control and Specimen Management in an International Multi-Center Investigation of Type 1 Diabetes: TEDDY

Table 1.

Blood Sampling Frequency and Volumes (all volumes are shown in milliliters)

HLA Screening at Clinical Centers with Confirmatory Testing at a Central Lab

The TEDDY cohort was enriched for subjects likely to reach disease endpoints by screening for specific HLA genotypes associated with moderate to high future risk of T1D. T1D incidence is relatively low compared to other chronic disease; therefore, TEDDY screened 424,788 children to identify its cohort. A total of 418,367 General Population (GP) infants were screened, of which 20,152 (4.8%) were eligible and 1,437 of the 6,421 screened infants (22.4%) with a First Degree Relative with T1D (FDR) were eligible. The details of this effort have been previously described [2]. The cohort was identified by HLA class II genotyping of newborn screening samples to allow determination of risk based on criteria established with pre-TEDDY data. The HLA typing was completed at five of the six international centers (the Finland Center conducted the typing for the German Center) using either asymmetric polymerase chain reaction and subsequent hybridization of allele-specific probes for HLA-DQA1, DQB1 and DRB1 as described [3] using DELFIA reagents (Perkin-Elmer, Waltham, MA USA) or a dot blot hybridization assay as detailed elsewhere [4]. Both methods used the same study inclusion criteria. There were separate inclusion criteria for the GP and FDRs of T1D patients. The FDR eligibility includes nine haplogenotypes (DR3/4, DR4/4, DR4/8, DR3/3, DR4/4b, DR4/1, DR4/13, DR4/9, and DR3/9) for broad HLA diversity, whereas the GP eligibility included only the first four haplogenotypes with DRB1*0403 as an exclusion allele. Errors in screening genotyping could originate from sample mislabeling, true genotyping errors, or rare haplotypes resulting in inferral errors. Central high-resolution confirmation testing was performed on all enrolled subjects and showed that the low-cost and low-resolution genotyping techniques employed at the screening centers yielded an accuracy of 99%. The TEDDY screening strategy demonstrated that different low-cost and low-resolution genotyping methods can result in efficient and accurate identification of a high-risk cohort for follow-up based on the TEDDY HLA inclusion criteria.

Genetic Proficiency Testing

The TEDDY study achieved excellent genotyping accuracy using genetic proficiency testing to ensure high initial and ongoing quality for T1D studies that employ HLA genetic risk assessment [5]. The Centers for Disease Control and Prevention (CDC) conducts both a voluntary quarterly proficiency testing (VQPT) program available to any laboratory and a mandatory annual proficiency testing (PT) challenge for TEDDY laboratories [5]. To mimic and test genotyping samples as those received by TEDDY, CDC sent whole blood and dried blood spots (DBS) samples with a wide range of validated HLA-DR and HLA-DQ genotypes to the five participating laboratories conducting screening tests and the centralized data center. Results were evaluated on the basis of both the reported haplotypes and the HLA genetic risk assessment. In the past six years, the VQPT reported from the 24 panels that 94.7% (857/905) of the relevant HLA-DR or HLA-DQ alleles were correctly identified with 96.4% (241/250) correctly categorized for risk assessment. There was significant improvement seen during the time interval of this program, with correct categorization reaching 100% during the last three years. TEDDY proficiency testing during the past four evaluations has revealed a genotyping accuracy of 99.9% (1153/1154). The different analytical methods used by T1D research centers have all provided accurate (>99%) results for genetic risk assessment. The two complementary CDC PT programs have documented the validity of the various approaches for screening and contributed to overall quality assurance.

Autoantibody Harmonization

Participants in TEDDY have autoantibodies measured starting at 3 months of age every 3 months until age 4 years, whereby based on appearance of a single persistent confirmed autoantibody the participant continues on the 3-month interval or if negative transitions to a 6-month interval until the age of 15 years (Table 1). As of May 2012, serum stored in the TEDDY Repository designated for autoantibody testing has been captured on 78% of the cohort adjusted for lost to follow-up (LTF) and withdrawn subjects. In an effort to ensure concordance between the two TEDDY core laboratories that process the autoantibody samples (Barbara Davis Center (BDC), Aurora, Colorado and the University of Bristol Laboratory, Bristol, UK), TEDDY participated in the NIDDK harmonization project in collaboration with the Diabetes Research Institute in Munich, Germany; the details have been published [6]. To evaluate the impact of the harmonized assay protocol on concordance of IA-2A and GADA results, two laboratories retested stored TEDDY Study sera using the harmonized assays. For IA-2A, using a common threshold of 5 DK units/ml, 549 of 550 control and patient samples were concordantly scored as positive or negative, specificity was greater than 99% with sensitivity 64% in all laboratories. For GADA, using thresholds equivalent to the 97th percentile of 974 control samples in each laboratory, 1051 (97.9%) of 1074 samples were concordant. On the retested TEDDY samples, discordance decreased from 4 to 1.8% for IA-2A (n=604 samples; P=0.02) and from 15.4 to 2.7% for GADA (n=515 samples; P< 0.0001). The precision of the measurement using the harmonized assay was determined in patient samples with values above 2.5 U/ml. The calibration of NIDDK calibrator samples and the median (interquartile range) of the coefficients of variation for IA-2A in the three laboratories (BDC, Bristol, and Munich) were 11.6% (10.6–21.5), 18.8% (15.8–23.9), and 8.7% (6.8–12.9); and for GADA were 10.8% (7.8–17.1), 9.1% (5.2–13.9), and 9% (6–15.3) [6]. The harmonization program for GADA and IA-2A was found to be feasible using large volume working calibrators and similar protocols. It provides a sound approach to ensure consistency in autoantibody measurements. In addition to the harmonized assay, the TEDDY study confirms all autoantibody positive samples and 5% of all negative samples at both laboratories. External quality control is also employed to test concordance by including external control samples identified and aliquoted in Gainesville, Florida and Munich, Germany, embedded in the routine monthly shipments to the two TEDDY core laboratories. These results are compared to external control reference results from the Munich laboratory.

Transglutaminase Autoantibody

Tissue transglutaminase autoantibodies (tGA) are serological markers for celiac disease, a chronic small bowel disease with autoimmune features caused by intolerance to dietary gluten and frequently detected among T1D patients. All children in TEDDY are annually screened for tGA starting at 24 months of age using radiobinding assays (RBA) analyzed in the same two core laboratories as the diabetes autoantibodies. If a child is positive for tGA at 24 months then all of their earlier samples (3 – 21 months) are tested for tGA. A child is classified as having persistent tGA when two consecutive measurements drawn at least three months apart are both positive. Children identified as persistent tGA in TEDDY are referred to a gastroenterologist for management at the clinical discretion of their primary provider. The decision to perform a biopsy is outside the TEDDY study protocol.

At the BDC, the RBA uses anti-IgA agarose to capture IgA-tGA; whereas, in Bristol a mixture of both anti-IgA agarose and protein A sepharose (PAS) is used to assess both IgA-tGA and IgG-tGA. Due to the discrepancy in the detection methods of the two IgA-tGA assays [7], TEDDY uses the BDC laboratory as a screen and the results are based upon the Bristol laboratory’s result. Children with a tGA level >0.05 at the BDC and >1.3 units (U) in Bristol are deemed tGA antibody-positive and are re-assayed in a follow-up sample taken after 3 months in samples collected at 24 and 36 months of age and after 6 months from 48 months of age. In order to exclude regional variations in the presence of tGA due to methodological differences between the BDC and Bristol laboratories, persistent tGA-positive sera were analyzed in a small set of samples at both laboratories for confirmatory analysis. In all, 900 TEDDY subjects were included for this exchange analysis and revealed a 93% concordance between the tGA assays, where 54 of discordant children were positive only in the Bristol and 3 in the BDC assays. Among those 54 children first detected in Bristol but not at the BDC, a total of 27 children reverted to tGA-negative after one or two follow-up samples, 17 remained persistent tGA-positive of which 10 children had subsequent follow-up samples with high tGA levels and were later diagnosed with celiac disease by their physician.

This finding resulted in a protocol modification in order to harmonize the tGA assays in TEDDY and it was decided that Bristol would be used as the reference laboratory in TEDDY for future analyses. The majority of discrepant samples were detected in individuals with low positive tGA levels in Bristol and had a tGA level between the detection level 0.01 and cut-off level 0.05 at the BDC. All sera with tGA levels >0.01 at the BDC are now re-assayed at the laboratory in Bristol for confirmation of persistence. Children found to be persistent positive in Bristol are not re-assayed at the BDC. In discordant cases, sera are further assayed separately for IgG-tGA by the PAS assay in order to distinguish between IgA-tGA and IgG-tGA positivity. Children with initial discordant sera later confirmed by the PAS assay are defined as persistent IgG-tGA positive.

Cryopreservation of Peripheral Blood Mononuclear Cells (PBMCs)

The PBMC sample collection was initiated in May 2010. PBMCs are isolated from TEDDY children at 3-month intervals using cell preparation tubes (CPT) and a freezing protocol derived from the immune tolerance network (ITN) guidelines [8,9]. To enhance isolated PBMC quality, technicians involved in cryopreservation of PBMCs are required to be annually certified. This certification procedure requires the collection of whole blood in two CPT from three separate non-TEDDY donors and the preservation of PBMCs in three separate vials frozen according to the TEDDY Manual of Operations (MOO). The samples are maintained at −80°C for two days before shipment to the central laboratory (Benaroya Research Institute, Clinical Core Lab, Seattle, USA) for quality control. Cells are evaluated for viability and recovery by certified technicians. The samples with documented sample volume and count information are shipped to the Benaroya Research Institute for centralized analysis. A passing score is determined by both cell recovery and viability with results reported by the DCC.

To optimize the recovery of PBMCs from CPT, a pilot study was conducted to define the best method for cell isolation and to minimize cell loss following centrifugation. Blood samples from 10 healthy donors were collected in three separate CPT and after centrifugation PBMCs were collected in three different ways: (1) removing the PMBCs layer without mixing the cells with the autologous plasma, (2) removing the cells after re-suspension in the autologous plasma by gentle pipetting, and (3) collecting cells as outlined above but the tubes were further washed twice with 2 ml of RPMI medium [9] to collect residual cells. Cell count and viability were determined by trypan blue exclusion assay with hemocytometer. The results demonstrated that a considerable amount of PBMCs remained in the CPT post-centrifugation and PBMC recovery was increased from 14% to 88% with inclusion of two additional washes of the CPT sample. Therefore, this improved protocol was added to the TEDDY MOO. As of May 2012, 45% of enrolled subjects have at least one preserved PBMC specimen.

Preservation of Enteroviruses in Stool Samples during Shipment

Stool samples are collected from all TEDDY children starting at 3 months of age for virological and other microbiological analyses. Samples are taken at home by the parents and sent to the TEDDY Repository by express mail (U.S. centers) or to the local repository by regular mail (European centers) where exposure to ambient temperatures for 1–3 days is possible. The effect of exposure to varying temperatures on enterovirus detection was evaluated by spiking enterovirus negative stool samples or sterile water samples with infective enterovirus (Coxsackievirus B3) and then exposing them to different temperatures for 8–72 hours. The varying temperatures evaluated were: 4.0°C/39.2°F, 25.0°C/77.0°F, 35.0°C/95.0°F, 43.0°C/109.4°F, 56°C/133°F and 65.0°C/149.0°F, at 8, 24, 48 or 72 hours.

The preservation of enterovirus RNA was first analyzed in spiked water samples by a sensitive semi-quantitative RT-PCR (5). The amount of viral RNA remained stable at temperatures ≤35°C/95°F for all tested time points. However, exposure to temperatures ≥43°C/109°F decreased the level of viral RNA detected, especially when exposed for 24 hours or longer. The highest temperatures (56°C/133°F and 65°C/149.0°F) rapidly decreased the amount of viral RNA even after the shortest (8 hour) exposure. The results were similar for the spiked stool samples.

The preservation of viral infectivity was analyzed by plaque assay (the number of plaque forming units in green monkey kidney (GMK) cells). Exposure to temperatures ≥56°C/133°F led to a rapid and complete loss of infectivity at 8 hours exposure. Exposure to 43.0°C/109.4°F for durations <72 hours reduced infectivity by 50–80% and durations ≥72 hours led to a complete loss of infectivity. Exposure to temperatures ≤25.0°C/77.0°F did not reduce infectivity even after 72 hours.

Overall, exposure to temperatures exceeding 35°C/95°F decreased both viral RNA and viral infectivity, and higher temperatures (56°C/133°F or higher) inhibited sensitive detection. Conversely, viral RNA and infectivity remained stable at lower temperatures. These findings resulted in the use of ice-gel stool sample shipments during the summer months to eliminate deleterious temperature peaks. The maximal temperature to which the samples were exposed during regular ice-gel packed shipments was monitored using commercially available CelsiStrip® temperature recording labels. Change in the color indicated the maximal temperature of exposure. These stickers were placed inside the stool sample mailing box during summer months in 1,467 shipments; the majority of samples (99%) remained at a temperature <40°C/105°F.

Monitoring Cross-contamination in TEDDY Blood and Stool Samples

This quality control protocol regularly monitors possible contamination of virus-negative samples by virus-positive samples (cross-contamination). This can happen when many samples are processed at the same time in the same location (e.g., when sample aliquots are made at TEDDY sites). Per protocol sterile phosphate buffered saline (PBS) samples are processed on a monthly basis in the same location as clinical samples. Possible contamination is then monitored by detecting enterovirus and rotavirus in these samples using sensitive RT-PCR. These viruses are common in children and are therefore optimal markers for contamination risk. PBS samples are processed in all sites which process blood, serum, plasma or stool samples in open tubes (e.g., dividing samples into smaller aliquots by pipettes or processing stool samples from diapers). PBS is first poured from a 500 ml bottle into a 50 ml tube in the same table or laminar flow where the samples are processed, and divided further into ten aliquots using the same procedures and equipment as for clinical samples. Samples are then shipped to the TEDDY Repository and stored at −80°C. Samples are analyzed regularly for the presence of rotavirus and enterovirus at the University of Tampere, Medical School (Tampere, Finland) using established RT-PCR methods validated in external quality control rounds (QCMD; http://www.qcmd.org or equivalent). If needed, the samples can also be screened for other infectious agents. To date, 492 samples have been tested, and all have been virus-negative.

Viral and Pathogen Detection Using Nasal Swabs

This pilot study evaluated the validity of identifying respiratory viruses and other pathogens using nasal swabs. Previous studies have found that nasal swabs are as effective as throat swabs for collecting samples from subjects with active upper respiratory infection; however, these methods have not been applied in regular sampling of asymptomatic subjects. Nasal swab sampling may be a non-invasive way to survey agents frequently causing infection and fever during childhood and the prevalence in the community during defined time periods.

Samples were collected using Copan nasal swabs from 50 healthy subjects (age 1.5 – 14 years, 23 March – 1 April 2009) and 40 subjects with active upper respiratory tract infection (age 0.5 years – adults, 28 May – 5 November 2009). All subjects were from Turku, Finland. Laboratory analyses were performed using virus specific PCR assays (RT-PCR) and multiple pathogen detection (multiplex PCR). For purposes of comparison, RT-PCR was performed at two independent sites in Finland (National Institute of Health and Welfare in Finland and University of Tampere, Finland). Multiplex PCR was performed at a third independent laboratory (Columbia University, New York) [13]. Results from all laboratories confirmed that respiratory pathogens can be detected using nasal swab procedures. The most frequently identified pathogens were human rhinovirus (HRV), H. influenzae, and S. pneumoniae. Viral pathogens were more frequently reported in subjects with active upper respiratory infection than in healthy individuals (HRV: 26/9 vs. H. influenzae: 15/2, S. pneumoniae: 2/6). Test accuracy was evaluated by false positive rates (FPR) and false negative rates (FNR). Using single pathogen detection in at least one lab to test subjects for clinical infections resulted in a FPR as high as 46/100 and a FNR of at least 22/100. However, estimating the frequency of asymptomatic and infectious subjects based on detection of multiple pathogens resulted in confirmed pathogen detection in 30/100 subjects with a low FPR of 6/100. Comparison of results across all three laboratories showed a high level of concordance with respect to HRV results. Confirmed positive HRV was detected in 65/100 symptomatic subjects with a FPR of 14/100. The virus specific PCR assay had an overall concordance rate of 98/100 for HRV. This assessment confirms that viral and bacterial pathogens can be detected using nasal swabs in both symptomatic and asymptomatic subjects.

Beginning at 9 months of age a minimally invasive nasal swab sample is collected from each TEDDY subject and continues to be collected on the visit schedule as outlined under Study Design and Population. Nasal swab collection began in December 2008 and 67% of the enrolled cohort have available samples for analysis.