CNVs are thought to significantly contribute to population-based adaptive evolution due to variability in expression levels. . Especially recently emerged genes may have significantly contributed to evolutionary change and phenotypic adaptation in more recently diverged evolutionary lineages . However, despite intensive work in the field, most CNVs and new resulting genes are so far, overall, rather poorly characterized at the functional level.
In this study, we provide a detailed functional analysis of a novel 56-kb deletion CNV involving two primate-specific genes of the BTNL family, namely BTNL8 and BTNL3, who share 80% homology in their coding sequence. The CNV has arisen in two 1.6-kb SD blocks of 98% identity, suggesting that the deletion occurred due to a NAHR event. CNVs arisen from NAHR tend to recur due to the unstable nature of the large and highly identical SD, leading to variants of different haplotypes that share common endpoints. Multiple genomic disorders as Williams-Beuren syndrome, Charcot-Marie Tooth disease, schizophrenia, and developmental delay have been assigned to NAHR events between misaligned LCRs [18–21], making NAHR the most common mechanism mediating recurrent microdeletions and microduplications.
By carefully looking at SNPs data of the International HapMap project and through our own genotyping data, we found currently existing frequency-matched SNPs in LD (r2 > 0.8) with the BTNL8*3 CNV in Eastern Asian, American and northern European populations. In contrast, no LD could be detected in African, Italian and Spanish populations. Especially African populations have been shown to be highly heterogeneous and to significantly vary in their haplotype structure and LD from other populations . However, also southern European subgroups as Italian, Spanish and Greek, have been shown to differ in their haplotype composition from northern and western European populations . This most likely results from large and constant migratory influences and admixture with other European and North African populations. These substantial differences between northern and southern European subgroups imply that many variations found through genome-wide association studies (GWAS) studies carried out on northern and western European population may not be replicated in southern European subgroups. Moreover, the absence of a tagging SNP in some populations has important implications for the interpretation of association studies. It stresses that unless a direct genotyping assay is applied for the BTNL8*3 CNV, its potential phenotypic impact may be overlooked in GWAS.
The BTNL8_BTNL3-del allele has been found at higher frequencies in European, East Asian and American populations, as compared to African, Oceanic and Middle Eastern populations. Ethnicity plays an important role in inter-individual variability of the immune system. Through recurrent exposure to different pathogens, ethnic groups have selected genetic adaptations that provide resistance or reduced susceptibility to infection  meaning that many times CNVs result in an advantageous phenotype for some populations [25, 26]. An example of a reduction in copy number being beneficial has been suggested for the α-globin locus, where it increases resistance to malaria infection and susceptibility to mild α-thalassemia. In regions where malaria is endemic, the deleted form of the α-globin gene can be found at an unusually high degree [27, 28]. Other examples where the number of gene copies positively correlates with infection are FCGR3B and DEFB4 genes, which are associated with glomerulonephritis, and Crohn’s disease, respectively [29, 30]
Here we show that, in accordance with the partial deletion of these genes, the BTNL8 and BTNL3 proteins were not expressed in LCLs homozygous for the deletion, and they were found at a significantly reduced amount in LCLs carrying one deleted allele. RT-PCR analysis, sequencing and western blot analysis revealed a new BTNL8*3 fusion product without any alterations in the coding frame, suggesting the existence of a novel, functional protein. In addition, BTNL9, another member of the butyrophilin family, was down-regulated in LCLs carrying the BTNL8_BTNL3-del allele. However, the function of BTNL8, BTNL3 and BTNL9 is yet unknown; therefore, here we can only speculate what might be the consequence of the deletion.
The BTNL proteins belong to the butyrophilin (BTN) family. The eponymous BTN protein (BTN1A1) is a type I transmembrane glycoprotein whose expression is restricted to the mammary gland during lactation, where it is involved in the secretion of milk fat globules . All BTN family members contain a signal peptide at their N-terminus, two Ig-like-domains (IgV and IgC) and a transmembrane-domain. The extracellular domain shows structural similarities to those of the B7 family, a protein family of co-stimulatory molecules involved in T cell activation . In contrast to B7 proteins, most BTN members harbour a cytoplasmic B30.2 domain at their C-terminus. Like the B7 protein family, several human and murine BTN and BTNL family members have been shown to control T cell response . While the B7 family of ligands and their receptors can regulate T cell response either through their positive (e.g. B7-1, B7-2, ICOS) or negative (e.g. PD-L1, PD-L2, B7-H3, B7-H4) co-stimulatory molecules, BTNs so far only have been found to act through co-inhibition [34–37]. In addition, polymorphisms in the human gene encoding for BTNL2 have been linked to a growing number of inflammatory diseases, e.g. sarcoidosis, myositis and inflammatory bowel disease [38–40]. Moreover, human BTN2A1 has been shown to modulate immature dendritic cells (DC) by binding to Dendritic Cell-Specific Intercellular adhesion molecule-3-Grabbing Non-integrin (DC-SIGN) , and Btnl1 has been recently found to regulate interactions with intraepithelial γδT lymphocytes in the murine small intestine by suppressing pro-inflammatory mediators of the NFκB pathway, such as IL-6, IL-15, CXCL1, and CCL4 . In addition, many members of the B7-homolog (B7-H) family, such as PD-1 and CTLA-4, are expressed on tumour cells in various cancers, where they can be exploited by the cancerous cells to escape from immune destruction and impede B7 ongoing immune processes [43, 44]. The presence of the deletion and the subsequent absence or reduced expression of the encoded proteins would allow a stronger response against tumour cells, implying that the BTNL8_BTNL3-del allele could act as a positive modulator of anti-tumour immunity. However, co-inhibitory molecules such as CTLA-4, PD-1 and BTLA have been shown to be crucial for the prevention of autoimmunity and polymorphism or deficiency of these molecules are associated with genetic susceptibility to autoimmune diseases in human and mice . Follow-up studies on the BTNL proteins will shed more light on their function in autoimmune diseases and cancer.
BTNL8 and BTNL3 are primate-specific genes, while BTNL9 has a clear ortholog in mice. Human BTNL8 and BTNL3 expression is primarily restricted to tissues of the digestive tract and at a lower level to spleen, thymus and lung. In addition BTNL3 is expressed in neutrophils and BTNL8 in eosinophils and at a reduced amount in neutrophils. In contrast, BTNL9 is mainly expressed in B cells and lymphoid organs, e.g. thymus, spleen, bone marrow and lymph nodes [46, 47]. Due to the high sequence homology and the similar expression-profile of BTNL8 and BTNL3 it is possible that the new BTNL8*3 fusion-protein compensates for the BTNL8 and BTNL3 wild-type proteins. However, more information is needed about the functions of BTNL8, BTNL3 and BTNL8*3 and the pathways they are involved in. Since almost all members of the BTN and BTNL families are highly expressed in the intestine , it has been postulated that these proteins act through a combined immunosuppressive effect, rather than a big impact of individual molecules . Therefore, it would be interesting to check for the consequences of the CNV in diseases associated with polymorphisms in BTN and BNTL genes, e.g. BTNL2 in Crohn’s disease or ulcerative colitis. Nevertheless, a review of the published GWAS data in these two disorders did not reveal an association with the potential tagging SNPs for the deleted allele (rs2387715, rs4700772 or rs10063135), although these SNPs are present in the affy 6.0 (rs2387715) and Illumina Omni 1 and human 1 M arrays (rs4700772 and rs10063135. There is, nevertheless, a replicated association on Chron’s disease to the 5q35 region around CPEB4 gene [49, 50], about 3 Mb away from the deletion.
Interestingly, when investigating gene expression changes with regard to the BTNL8_BTNL3-del allele, we found TNF and the ERK1/AKT pathway to be central hubs of the network influenced by the deletion CNV in LCLs. TNF, ERK1 and AKT are important players in signal transduction pathways and key components of the immune response in humans and even a slight deregulation of those proteins might have an important impact in the response to pathogens. However, even though the HapMap repository represents a fantastic source for genetic studies, the analysis was limited due to the use of LCLs, which might not be the main cell-type where the BTNL8*3 CNV affects expression levels, since BTNL3 and BTNL8 are predominantly expressed in the digestive tract. Follow-up studies using other cell types will be needed.