Advances in genetic technology have rapidly changed healthcare delivery in low- and middle-income countries. NGS utilization has decreased the time to diagnosis, increased the diagnostic rate, and provided valuable insight into the genotype–phenotype correlation of IEI in a timely and cost-effective way28,29. IEI is not uncommon in India; However, their diagnosis is either missed or delayed due to a lack of awareness and a paucity of diagnostic facilities. There is an urgent need to increase testing capacity for early recognition, diagnosis, and management of IEI in our country30,31,32.
We have been diagnosing patients with IEI at our center for the past 25 years. However, services for molecular diagnosis for IEI both in government and commercial sectors have not been available in India until 2016. For molecular diagnosis of IEI, we established academic collaboration with Service Hématologie, Immunologie et de Cytogénétique, Hôpital de Bicêtre, Le Kremlin Bicêtre, at France in the year 2007. Later, we established collaboration with institutes at Japan (National Defense Medical College, Saitama) and Hong Kong (Department of Pediatric and Adolescent Medicine, University of Hong Kong) in the years 2008 and 2010, respectively. This has facilitated molecular diagnosis for many of our patients with IEI. Our center was designated as Center for Advanced Research in diagnosis and treatment for primary immunodeficiency diseases by the Indian Council of Medical Research, Government of India, in 2015. Until 2016, tests available for diagnosis of IEI at our center include immunoglobulin estimation, NBT, and flow cytometry for several surface and intracellular proteins10. With the increase in patients diagnosed with IEI in the last few years, we felt the need to establish molecular analysis at our center4. We initiated Sanger sequencing for BTK, CYBBand WAS genes in our center in 2016 (Fig. 1).
Commercial laboratories in India came up with facilities (targeted exome) for molecular diagnosis of IEI in 2016. Costs incurred for sequencing in commercial laboratories were exorbitant (USD 400–500) in 2016 that later reduced in the subsequent years (USD 200 currently). The introduction of targeted NGS for IEI in 2018 at our center has enabled us to offer this diagnostic modality to many of our patients who could not afford the costs of commercial testing. We have also been able to diagnose more IEIs each year and at a much faster pace than in previous years. The cost of targeted genetic sequencing at our setup is USD 83 per sample. This is much less than the costs incurred at commercial laboratories in India33. In addition, infants less than one year are covered under the JSSK (Janani Sishu Suraksha Karyakram) scheme of the Government of India. They are entitled to avail of NGS free of cost. Our Institute also provides free diagnostic services to patients from low-income groups who cannot afford the NGS charges, and charges are minimal for those who can afford this facility.
We have worked upon and improvised the standard protocol of NGS to suit our setup. We made some ingenious modifications to the recommended protocol to reduce the cost per sample and accommodate more patient samples in each run. Towards this end, we have successfully used half the recommended volume of reagents (however, concentration remained the same) at each successive step by starting with an initial DNA volume of 2.5µL instead of 5µL. So, a larger number of patient samples could be accommodated in each run. We have effectively run 42 patient samples with a 24-reaction reagent kit for 24 samples.
NGS sample preparation is a tedious and labor-intensive process requiring focus and concentration at each successive step34,35. After chip-loading and sequencing, we did not get results for two runs. On both these occasions, instead of repeating from the start, we started after the library quantification step as we were sure about the quality of the library preparation. So, restarting with the template preparation step instead of beginning from the start in the case of a failed run could be a helpful strategy if we are sure about the quality of library preparation.
We describe preliminary results of targeted NGS in 121 patients with different forms of IEIs diagnosed and managed at our centre. Our variant pick-up rate of 63.6% is much higher than previous studies- 25% by Yska et al. in 2019 and 29% by Vorsteveld et al. in 202128,36. The pick-up rate of variants in other studies was 16%7 (Gallo et al., Italy, 2016), 14% (Kojima et al., Japan, 2016)372.1% (Sun et al., China in a cohort of infants)3828.6% (Cifaldi et al., Italy, 2020)18 and 42.4% (Arunachalam et al., India, 2020)33.
There are several reasons for a higher diagnostic yield in our study. Careful patient selection with a high pre-test probability based on clinical manifestations and preliminary immunological investigations was done. Patients with a high probability of having a pathogenic variant in one of 44 genes included in the gene panel are sorted out in consultation with clinicians trained in immunology and have broad experience in caring and managing patients with IEI. Currently more than 400 genes are implicated in various IEI. However, we selected 44 genes based on the most common diseases we encounter at our center and also since we aim to provide genetic diagnosis to a maximum number of patients at an affordable cost. A large panel although more desirable would be costlier to design and in addition fewer samples would be accommodated in each run. Samples of patients who are very likely to have genetic variants in the genes included in the panel were included based on clinical history and initial immunological investigations. Patients with IEI not clearly delineated upon initial immunological investigations are referred for a clinical exome or whole-exome analysis. This analysis is outsourced to commercial laboratories providing these services at an affordable cost.
NGS has facilitated the early diagnosis of patients with IEI in situations where flow cytometry was either not conclusive or did not match the clinical presentation. For instance, patient 56 was clinically suspected of having an autosomal recessive hyper-IgM was found to have biallelic variants in the ATME gene. Hence, relying on typical manifestations of the IEI may not be ideal, and a rapid genetic diagnosis is indispensable39.
There have also been instances when the initial analysis on the Ion Reporter did not reveal a pathogenic variant. In patient 8 with clinically suspected XLA, no pathogenic variant was detected at initial analysis. There was a strong clinical suspicion of XLA in this case; we manually visualized the data on Integrative Genomics Viewer (IGV). We found a large deletion of exon-10, 11 and 12 in the BTK gene (Fig. 2)40. Similarly, in another suspected patient with CGD (Pt.27), a large deletion was found in the CYBA gene, which was missed by the ion reporter software but was detected on manual reanalysis and visualization on the IGV. Patient 42 had an indel in IL2RG gene. In patient 42, analysis by the Ion reporter software revealed two IL2RG variants in close proximity, which was confusing. However, upon visualization of the BAM file on IGV, we realized that it was an indel (insertion of 3 nucleotides and deletion of 8 nucleotides) which was misinterpreted as two variants by the ion Reporter software.
Hence, manual data visualization on IGV and manual analysis of annotated vcf files instead of relying on variants detected by initial analysis by software is crucial. We have been able to detect these variants in these cases using this strategy.
Detection of genetic variants in genes with known pseudogene is another problem that we encountered in our patient cohort. We faced this difficulty in patients with autosomal recessive CGD due to NCF1 gene defect. The targeted NGS panel systematically missed the most common pathogenic variant in NCF1ie, deletion of two nucleotides at the start of Exon-2. NCF1 gene has two flanking pseudogenes (ΨNCF1)41. We assume that the amplicon designed for exon-2 of the NCF1 gene was unable to bind to its target, and thus, there was no amplification of this region, resulting in no reads for exon-2 in these patients. We performed a gene scan in 3 patients who had no reads in Exon-2 of the NCF1 gene to check for this variant and confirmed NCF1 GT deletion in all 3 of these patients (Fig. 3A,B).
We have also been able to offer prenatal services to many patients. Patient 40 was clinically suspected of having SCID but had expired before a genetic defect could be established. His mother was pregnant at this time, and the period of gestation was 13 weeks. We were able to identify a splice-site variant in the IL2RG gene in this family with X-linked SCID, and the mother was offered prenatal diagnosis by chorionic villous sampling. Molecular confirmation of diagnosis helped the family to get timely antinatal testing and appropriate genetic counseling. For some patients, especially SCID, rapid diagnosis through targeted NGS has saved lives, or genetic counseling has prevented an affected child in the subsequent pregnancy.
Pt the mother of a diagnosis 76 dead child was suspected to have X-linked Hyper-IgM, but a genetic could not be established during the child’s life. Targeted NGS revealed a synonymous variant in exon 1 of the CD40LG gene proximal to donor splice-site. In-silico prediction for this variant was found to be ‘damaging’ by Mutation Tester2. Synonymous variants involving canonical splice-sites can also be pathogenic and should not be filtered out.
Genetic findings were beneficial in providing genetic counseling to affected families, carrier screening, and prenatal diagnosis. Moreover, genetic information is required for devising appropriate transplantation related strategies. Genetic findings were also crucial in deciding the treatment modalities in a few cases. Cases harboring defects leading to antibody deficiencies were placed on regular replacement intravenous immunoglobulin therapy.