BBLN triggers CAMK2D pathology in mice under cardiac pressure overload and potentially in unrepaired hearts with tetralogy of Fallot

Tetralogy of Fallot (TOF) is one of the most prevalent congenital heart defects, with adverse cardiac remodeling and long-term cardiac complications. Here, searching for pathomechanisms, we find upregulated bublin coiled-coil protein (BBLN) in heart specimens of TOF patients with cyanosis, which positively correlates with cardiac remodeling pathways. Human BBLN, a protein with largely unknown function, promoted heart failure features, with increased mortality when overexpressed in mice, in a protein dosage-dependent manner. BBLN enhanced cardiac inflammation, fibrosis and necroptosis by calcium/calmodulin-dependent protein kinase II delta (CAMK2D) activation, whereas a BBLN mutant with impaired CAMK2D binding was inert. Downregulation of CAMK2D by an interfering RNA retarded BBLN-induced symptoms of heart failure. Endogenous BBLN was induced by hypoxia as a major TOF feature in human patients and by chronic pressure overload in mice, and its downregulation decreased CAMK2D hyperactivity, necroptosis and cardiovascular dysfunction. Thus, BBLN promotes CAMK2D-induced pathways to pathological cardiac remodeling, which are triggered by hypoxia in TOF. Abd Alla et al. identify bublin coiled-coil protein (BBLN) as a factor upregulated in unrepaired hearts of patients with tetralogy of Fallot, and show that this protein binds to and controls the function of calcium/calmodulin-dependent protein kinase II delta (CAMK2D). Overexpression of BBLN in mice induced pathological cardiac remodeling, which was precluded when the interaction of BBLN with CAMK2D was prevented or upon CAMK2D downregulation.

Tetralogy of Fallot (TOF) is one of the most prevalent congenital heart defects, with adverse cardiac remodeling and long-term cardiac complications.Here, searching for pathomechanisms, we find upregulated bublin coiled-coil protein (BBLN) in heart specimens of TOF patients with cyanosis, which positively correlates with cardiac remodeling pathways.Human BBLN, a protein with largely unknown function, promoted heart failure features, with increased mortality when overexpressed in mice, in a protein dosage-dependent manner.BBLN enhanced cardiac inflammation, fibrosis and necroptosis by calcium/calmodulin-dependent protein kinase II delta (CAMK2D) activation, whereas a BBLN mutant with impaired CAMK2D binding was inert.Downregulation of CAMK2D by an interfering RNA retarded BBLN-induced symptoms of heart failure.Endogenous BBLN was induced by hypoxia as a major TOF feature in human patients and by chronic pressure overload in mice, and its downregulation decreased CAMK2D hyperactivity, necroptosis and cardiovascular dysfunction.Thus, BBLN promotes CAMK2D-induced pathways to pathological cardiac remodeling, which are triggered by hypoxia in TOF.
It is estimated that over 7,000 rare diseases affect about 300 million people worldwide 1 .By definition, a rare disease affects fewer than 5 out of 10,000 people, and so far, the causes of most rare diseases are not understood 1 .On the basis of many successful examples, the elucidation of rare diseases is not only relevant for the few directly affected patients but also supports precision medicine approaches for patients with similar common pathologies 2,3 .With a worldwide prevalence of 3-4 cases per 10,000 live births, tetralogy of Fallot (TOF) is also a rare disease.In addition, TOF is among the most prevalent congenital heart defects and the most frequent congenital cyanotic heart defect [4][5][6] .
It is characterized by four major cardiac abnormalities, which consist of a ventricular septal defect, an overriding aortic root, an infundibular stenosis of the pulmonary artery and a right ventricular hypertrophy [4][5][6] .Heart defects of TOF cause impaired oxygen delivery to the infant and, depending on severity, culminate in cyanosis [4][5][6] .Treatment of TOF relies on surgery to correct cardiac defects 7 .Despite greatly improved surgical procedures in infancy and early childhood, long-term complications of corrected TOF remain a major problem [7][8][9][10] .Besides coronary artery disease, long-term cardiac complications of TOF include cardiac hypertrophy and cardiac remodeling, which predispose to an increased Article https://doi.org/10.1038/s44161-023-00351-6because the pathomechanisms underlying the cardiac remodeling process triggered by TOF are unknown [7][8][9][10] .
In this Article, to identify pathomechanisms triggered by TOF, we searched for CAMK2D-interacting proteins in cardiac specimens of TOF patients undergoing surgical correction of the right ventricular outflow tract (RVOT) obstruction.
We identified bublin coiled-coil protein (BBLN) as a CAMK2Dinteracting protein that is upregulated in TOF cyanotic patients compared with noncyanotic ones.Given the largely unknown function of BBLN, we generated two transgenic mouse models overexpressing the gene 25 or 50 times more than wild-type mice, and found that BBLN overexpression induced cardiac dysfunction in mice in a dose-dependent way.Downregulation of CAMK2D by an interfering RNA retarded BBLN-induced symptoms of heart failure.We found that endogenous BBLN was induced by chronic pressure overload in mice, and it is probably induced by hypoxia in cyanotic TOF patients.BBLN downregulation in mice exposed to pressure overload, decreased CAMK2D-induced inflammation, fibrotic remodeling and necroptosis.Overall, our study suggests BBLN as a regulator of CAMK2D activity and a potential target to prevent adverse cardiac remodeling in different conditions, including TOF.

BBLN is upregulated in TOF patients with cyanosis
Immunoaffinity enrichment (AP) of CAMK2D from TOF patient heart specimens, and nanoliquid chromatography-electrospray ionization-tandem mass spectrometry (nano-LC-ESI-MS/MS) identification of co-enriched proteins in the low molecular weight 8-14 kDa range, revealed BBLN as a previously unrecognized CAMK2D-interacting protein (Fig. 1a).Immunoblot (IB) detection confirmed the identity of the CAMK2D-interacting protein as BBLN (Fig. 1b).Immunohistological analyses found that BBLN protein contents were increased sixfold (range 1.9-to 11.45-fold) on cardiac biopsy specimens of TOF patients with cyanosis compared with those of acyanotic TOF cases (Fig. 1c-g).
The human BBLN gene is an early growth response 1 (EGR1)responsive gene with an EGR1 consensus site in the promoter region, and according to Gene Expression Omnibus (GEO) dataset GDS2008, BBLN is induced by EGR1 expression (Extended Data Fig. 1a,b).The upregulation of the human BBLN protein in cyanotic TOF patient hearts could thus be a direct consequence of the low oxygen condition because EGR1 is upregulated under hypoxic conditions 14 .In agreement with this notion, ventricular specimens from cyanotic TOF patients displayed significantly upregulated EGR1 transcript levels compared with acyanotic TOF cases (data from ref. 15 are shown in Extended Data Fig. 1c).Although the human and murine BBLN amino acid sequences have 94% identity, the murine Bbln gene lacks the Egr1 consensus site (Extended Data Fig. 1a,d) and could therefore be upregulated by other modalities in different heart failure models (Extended Data Fig. 1e,f).

Tg-BBLN mice developed dilative cardiac hypertrophy
As the in vivo function of the human BBLN protein is largely unknown, we investigated the impact of an increased cardiac BBLN content by generation of BBLN-transgenic (Tg-BBLN) mice.Although TOF originates in the right ventricle, the BBLN-inducing hypoxia in cyanotic TOF patients affects the whole heart, and the deoxygenated blood flows through the right and left ventricle and the coronary blood vessels.Therefore, BBLN was expressed under control of the alpha-myosin heavy chain (MHC) promoter, which induces myocardium-specific BBLN expression in the whole heart (Fig. 2a).Two different Tg-BBLN mouse lines were generated (Tg-1 and Tg-2) with 50.4 ± 8.2-fold and 25.1 ± 9.0-fold increased cardiac BBLN protein contents, respectively, compared with those of nontransgenic Friend leukemia virus B susceptible (FVB) control mice (Fig. 2b,c), which feature the increased BBLN contents of TOF patient hearts (cf.Fig. 1e).Immunohistological analysis of hearts from 3-4-month-old Tg-BBLN mice confirmed increased cardiac BBLN protein contents and showed cardiomegaly with enlargement of right and left ventricles (Fig. 2d).Frequently, Tg-BBLN mice (Tg-1) developed atrial myxomas (Fig. 2d), which are also reported in some TOF cases 16 and could arise from multipotent cardiac stem cells during unsuccessful heart repair 17 .However, increased BBLN contents in cardiac specimens of TOF patients were a consequence and not a cause of TOF defects because Tg-BBLN mice with heart failure had no septal defect (Fig. 2d).Concomitant with cardiac enlargement, the increased cardiac BBLN protein content triggered cardiac dysfunction (Fig. 2e).The cardiac phenotype of Tg-BBLN mice was BBLN protein dosage dependent.Tg-BBLN mice with 50.4 ± 8.2-fold increased cardiac BBLN protein levels (Tg-1), had an early heart failure phenotype with a strongly reduced left ventricular ejection fraction (EF) of 18.0 ± 6.0% and an increased mortality starting at an age of 3 months (Fig. 2e,g).
At the same age, Tg-BBLN mice with 25.1 ± 9.0-fold increased cardiac BBLN levels (Tg-2) had a significantly decreased left ventricular EF of 33.0 ± 5.8% compared with 52.3 ± 4.2% of age-matched, nontransgenic FVB mice (Fig. 2e).The 1-year survival rate was 19.6% of male Tg-1 mice and 90.8% of male Tg-2 mice (Fig. 2g).The heart failure phenotype with an increased mortality was also observed in female Tg-BBLN mice (Extended Data Fig. 2a,b).Echocardiographic analyses confirmed the immunohistological assessment and showed that the heart failure phenotype of Tg-BBLN mice was accompanied by enlargement of the left cardiac ventricles (Fig. 2f and Extended Data Fig. 2c).

Tg-BBLN mice reproduce features of TOF patients
To investigate the relationship between increased cardiac BBLN contents and TOF-related gene expression, we performed a whole-transcriptome analysis of RVOT specimens from TOF patients with cyanosis (Extended Data Fig. 3) and right ventricular heart specimens from Tg-BBLN mice (Extended Data Fig. 4).Next-generation sequencing (NGS) transcriptome profiling of right ventricular heart tissue of Tg-BBLN mice was validated by the detection of right ventricle-specific transcript changes in comparison with left ventricular tissue (Extended Data Fig. 4).
Whole-transcriptome analysis found that about 60% of transcripts, which positively correlated with BBLN levels in TOF patient heart specimens with cyanosis (874 transcripts), also showed upregulation in the right ventricles of Tg-BBLN mice (Supplementary Dataset 1).The positive correlation of a transcript with cardiac BBLN transcript levels in TOF patient heart specimens is exemplified for the heart failure and myocardial ischemia marker, natriuretic peptide type b (NPPB) (Fig. 2h), which could be induced in TOF patients with cyanosis by the low oxygen condition 18,19   Nppb was also upregulated in right ventricular heart specimens of Tg-BBLN mice (Fig. 2i).The overrepresentation analysis of transcripts having positive correlation with BBLN levels in right heart specimens of TOF patients with cyanosis and showing upregulation in right ventricles of Tg-BBLN mice documented that BBLN triggered major biological pathways involved in heart remodeling (Figs.2j and 3a and Supplementary Dataset 1).These pathways included the immune system pathway, the extracellular matrix organization pathway and the programmed cell death pathway with transcripts regulating apoptosis and necroptosis (Figs.2j and 3a and Supplementary Dataset 1).
The immune system pathway was the major upregulated pathway triggered by BBLN in right TOF patient heart specimens and in right Tg-BBLN mouse hearts (Fig. 2j and Fig. 3a,b).Of those BBLN dependently upregulated cardiac transcripts of the 'immune system' pathway of TOF patients and Tg-BBLN mice, 23.8% have a documented pathological relevance for cardiac inflammation, cardiac remodeling and/or heart failure (Fig. 3b,c).The transcript intensity heat map documents the predominant upregulation of the heart failure-enhancing transcripts in right ventricles of Tg-BBLN mice compared with left cardiac ventricles (Fig. 3c).Thus, BBLN triggered the predominant upregulation of heart failure-enhancing transcripts in the right ventricles of Tg-BBLN mice and TOF patients.
The transcriptome analysis also found that 114 transcripts showed negative correlation with BBLN in TOF patients with cyanosis and downregulation in right ventricular heart specimens of Tg-BBLN mice (Supplementary Dataset 2).The overrepresentation analysis of those 114 transcripts displaying negative correlation with BBLN in TOF patient hearts and downregulation in Tg-BBLN mice revealed that increased BBLN led to downregulation of major cardiac metabolism pathways, such as the citric acid cycle and respiratory electron transport pathway, the mitochondrial translation pathway and the branched-chain amino acid catabolism pathway (Fig. 2j).Reduced mitochondrial respiratory chain activity, defective mitochondrial protein translation and impaired branched-chain amino acid catabolism are prominent features of patients with TOF and/or other forms of cardiomyopathy and heart failure [20][21][22] .Taken together, the right ventricular heart tissue of transgenic Tg-BBLN mice with increased cardiac BBLN contents reproduced major cardiac gene expression changes and alterations of biological pathways, which are characteristic of right ventricular TOF patient heart specimens.

BBLN promoted CAMK2D activation in vivo
We searched for molecular mechanisms evoked by BBLN in the heart because the function of the human BBLN protein is largely unknown.BBLN was identified as a CAMK2D-interacting protein in TOF patient hearts (cf.Fig. 1).Immunofluorescence detected colocalization of BBLN with CAMK2D in areas with cardiac degeneration in Tg-BBLN heart specimens (Fig. 4a and Supplementary Fig. 1).In addition, Tg-BBLN mice showed increased cardiac CAMK2D activity with significantly elevated levels of autophosphorylated phospho-T287-CAMK2D, which were 9.2 ± 0.5-fold higher in Tg-BBLN hearts than in nontransgenic FVB controls (Fig. 4b).Of note, autophosphorylation of CAMK2D on threonine 287 causes autonomous kinase activity in the absence of its stimuli Ca 2+ and calmodulin 23 .In contrast, the inactivating autophosphorylation of CAMK2D on T307 (and its neighbor T306), which occurs in the absence of calcium and prevents Ca 2+ /calmodulin binding 24 , was almost undetectable in Tg-BBLN hearts, whereas inactivating autophosphorylation on T307 was present in nontransgenic FVB hearts (Fig. 4b).For comparison, total CAMK2D protein levels of Tg-BBLN hearts were slightly but significantly (1.12-fold) above those of nontransgenic FVB controls (Fig. 4b).Likewise, cardiac transcript levels of the most abundant delta-C-related Camk2d splice variant 9 in the heart 12,25 and of the delta-C splice variant were slightly (1.12-and 1.32-fold) increased in right ventricles of Tg-BBLN hearts (Extended Data Fig. 5a).Notably, these transcripts were only increased in right ventricles but not in left ventricles of Tg-BBLN hearts (Extended Data Fig. 5a,e).
Transcript levels of the Camk2d delta-A splice variant were low and slightly higher in both right and left ventricles of Tg-BBLN mice   compared with those of nontransgenic FVB mice, whereas the delta-B splice variant was barely detectable (Extended Data Fig. 5a,e).Other Camk2 genes (Camk2a, Camk2b and Camk2g) showed low cardiac expression levels (Extended Data Fig. 5b-d,f-h).Taken together, BBLN increased the cardiac levels of activated phospho-T287-CAMK2D.Concomitantly, CAMK2D substrate phosphorylation was also enhanced in Tg-BBLN hearts as exemplified for the ryanodine receptor 2 (RYR2) phosphorylation on serine 2813 (Extended Data Fig. 6a).Tg-BBLN hearts had increased contents of serine 2813-phosphorylated RYR2 (Extended Data Fig. 6a), which could contribute to BBLN-induced symptoms of heart failure 26 .Together, these experiments showed that BBLN enhanced the CAMK2D activity in vivo, in Tg-BBLN mice.

BBLN enhanced CAMK2D activation in vitro
In vitro experiments confirmed a causal relationship between BBLN and CAMK2D activity enhancement (Fig. 4c,d).Recombinant purified BBLN protein augmented the activating CAMK2D autophosphorylation on T287 in vitro (Fig. 4c and Extended Data Fig. 7a).BBLN also enhanced the CAMK2D-mediated substrate phosphorylation as proved with phosducin (PDC) (Fig. 4c,d), which was used as a CAMK2 substrate 27 .In addition, CAMK2D also phosphorylated BBLN (Fig. 4c).The CAMK2D-enhancing effect of BBLN was linked to the presence of Ca 2+ /calmodulin because without addition of Ca 2+ /calmodulin, the CAMK2D-enhancing effect of BBLN was abolished, and the basal CAMK2D autophosphorylation and substrate phosphorylation were strongly reduced (Fig. 4c,d).
Searching for the mode of interaction between BBLN and CAMK2D, we performed sequence alignment of BBLN with the prototypical CAMK2-interacting peptide of glutamate ionotropic receptor NMDA type subunit 2B (GRIN2B) residues 1,295-1,310, and with the regulatory segment residues 275-295 of CAMK2D.The alignment showed that the human BBLN amino acid sequence displays major features, which are typical for CAMK2 kinase domain interaction and which are present in CAMK2 substrates (Extended Data Fig. 7b).Notably, major interaction sites of GRIN2B (residues 1,295-1,310) and of the regulatory segment residues 275-295 with the CAMK2 kinase domain 28,29 were present in BBLN residues 55-74 (Extended Data Fig. 7b).BBLN could thus interact with the previously identified binding site for activating substrates in the CAMK2(D) kinase domain 28 .In agreement with this notion, we found that the interaction of BBLN with the CAMK2D kinase domain was strongly reduced in the presence of the kinase domain-binding peptide of GRIN2B (Extended Data Fig. 7c).
We generated different BBLN mutants to analyze the interaction of BBLN with CAMK2D.The BBLN-SxxA mutant with removal of all phosphorylation sites (BBLN-S2A, Y28A, S33A, S40A, S62A, T66A, S79A and S82A) was not phosphorylated by CAMK2D (Extended Data Fig. 7d) and showed a strongly reduced interaction with the CAMK2D kinase domain, as determined by co-enrichment (Extended Data Fig. 7e).For comparison, BBLN-Mut1 (BBLN-S2A, T66A, S79A and S82A) and BBLN-Mut2 (Y28A, S33A and S40A) showed a partially reduced phosphorylation by CAMK2D (Extended Data Fig. 7d).However, the interaction with the CAMK2D kinase domain was still effective and comparable to wild-type BBLN (Extended Data Fig. 7e).This observation is in agreement with other CAMK2 substrates such as GRIN2B, in which exchange of a single CAMK2 phosphorylation site to alanine did not disrupt the binding to CAMK2 (ref.30).The BBLN-SxxA mutant with impaired interaction with the CAMK2D kinase domain also showed a strongly reduced CAMK2D-enhancing effect in vitro, whereas BBLN-Mut1 and BBLN-Mut2 were still able to enhance CAMK2D autophosphorylation on T287 (Extended Data Fig. 7a).Together, these in vitro data strongly indicate that BBLN functions similarly as other CAMK2-activating substrates 28 .BBLN does not induce the active 'on' state but could sustain/ enhance the activating CAMK2D autophosphorylation on T287, which is triggered by the presence of Ca 2+ /calmodulin.

BBLN triggered cardiac dysfunction by CAMK2D activation
In vivo experiments confirmed a major role of BBLN-stimulated CAMK2D activity in the pathologic cardiac phenotype of Tg-BBLN mice because downregulation of the CAMK2D protein by lentiviral transduction of a Camk2d-targeting micro (mi)RNA prevented the BBLN-induced increase in cardiac contents of autophosphorylated phospho-T287-CAMK2D, reduced the cardiac phosphorylation of the CAMK2D substrate RYR2 on S2813 and retarded the BBLN-induced cardiac dysfunction (Fig. 4e,f).Likewise, the BBLN-SxxA mutant, which was ineffective in vitro, did not enhance CAMK2D autophosphorylation and CAMK2D substrate phosphorylation in vivo, in transgenic Tg-BBLN-SxxA mice (Fig. 4e).Moreover, the inactive BBLN-SxxA mutant did not cause cardiac dysfunction and cardiac hypertrophy (Fig. 4f and Extended Data Fig. 8).Nevertheless, at present, we cannot completely rule out that CAMK2(D)-independent effects may contribute to the unaltered cardiac phenotype of Tg-BBLN-SxxA mice.
Similar to transgenic Tg-BBLN mouse hearts, cardiac specimens of human TOF patients with increased BBLN contents and cyanosis also showed increased phospho-T287-CAMK2D contents compared with those of acyanotic TOF patients, whereas total CAMK2D protein levels were not significantly different between the two groups (Extended Data Fig. 9a).Taken together, our data strongly suggest that BBLN triggered cardiac dysfunction and cardiac hypertrophy by CAMK2D activation.
Upregulation of key transcripts of cardiac fibrosis and heart remodeling such as Col1a1, Col1a2, Col3a1 and the collagen-cross-linking enzyme Loxl2 were also detected in Tg-BBLN hearts with only a four-fold increase in BBLN protein levels (Extended Data Fig. 10a-c).Moreover, these mice with moderately increased cardiac BBLN levels developed features of compensated heart failure at an age of 8 months, with a significantly reduced left ventricular EF of 40.5 ± 2.7% compared with 53.2 ± 4.5% of nontransgenic FVB mice, and showed a significant cardiac enlargement in echocardiographic examination (Extended Data Fig. 10a,b).Thus, moderately elevated cardiac BBLN levels are sufficient to cause fibrotic cardiac remodeling and features of heart failure.

BBLN-triggered calcium overload and DES degradation
Concomitant with necroptotic cell death, which is associated with calcium overload 37 , Tg-BBLN hearts were characterized by calcium overload, which was detected by von Kossa staining (Fig. 6g).Calcium accumulation was accompanied by disturbance of the intracellular calcium-handling proteins in Tg-BBLN hearts (Extended Data Fig. 6).Notably, the cardiac RYR2 and SERCA2 (ATP2A2) contents were downregulated at the protein and transcript level in Tg-BBLN hearts (Extended Data Fig. 6a-c).Likewise, in TOF patient hearts, the RYR2 and SERCA2 transcript levels were negatively correlated with BBLN (Extended Data Fig. 6d).With these features, Tg-BBLN mouse hearts and human heart specimens from TOF patients with cyanosis resemble patients with ischemic cardiomyopathy, whose hearts also show decreased RYR2 and SERCA2 levels 38,39 .
Calcium enhances the degradation of the major intermediate filament protein, desmin (DES), in heart failure 40,41 , and BBLN interacts  with intermediate filament proteins 42 .Tg-BBLN mice with signs of heart failure and calcium accumulation also showed DES fragmentation, in contrast, the DES protein was largely intact in transgenic Tg-BBLN-SxxA mice with expression of the inactive BBLN-SxxA mutant (Fig. 7a).BBLN-induced DES fragmentation was related to BBLN-enhanced CAMK2D activation, because downregulation of CAMK2D protein levels by miCamk2d retarded the BBLN-induced DES fragmentation (Fig. 7b).Together, these data provide strong evidence that BBLN causes features of cardiac degeneration by direct activation of CAMK2D.

Upregulation of BBLN by chronic pressure overload in mice
Human BBLN is an EGR1-responsive gene, which is upregulated in TOF patient heart specimens with cyanosis (Extended Data Fig. 1a-c and ref. 15).However, the EGR1 response element is not present in the mouse Bbln gene (Extended Data Fig. 1a).Nevertheless, previous studies showed that the murine and rat Bbln were triggered in different heart failure models (Extended Data Fig. 1e,f and murine BBLN amino acid sequences have 94% identity (Extended Data Fig. 1d).Endogenous Bbln upregulation was induced in a murine heart failure model, which was triggered by chronic pressure overload imposed by abdominal aortic constriction (AAC) (Fig. 8).Two months of AAC led to a significant increase in cardiac murine BBLN protein contents together with increased activated phospho-T287-CAMK2D levels (Fig. 8a).Concomitantly, the necroptosis effector, that is, the aggregated phospho-S345-MLKL octamer, was triggered (Fig. 8b), and showed colocalization with BBLN on cardiac specimens with necroptosis (Fig. 8c).Downregulation of endogenously expressed Bbln by transgenic expression of a Bbln-targeting miRNA under control of the ubiquitous cytomegalovirus (CMV) promoter led to decreased cardiac phospho-T287-CAMK2D contents, reduced the levels of aggregated phospho-S345-MLKL octamers as effectors of necroptosis and improved the cardiac function in heart failure induced by AAC (Fig. 8a-d).Thus, chronic pressure overload promoted the upregulation of the endogenous BBLN protein in the mouse.These pathologically elevated endogenous cardiac BBLN protein levels contributed to cardiac dysfunction and necroptotic cardiac remodeling in vivo, in the AAC-induced heart failure model.
The in vivo function of the human BBLN protein, which is encoded by the chromosome 9 open reading frame 16 (C9orf16; ref. 43) was largely unknown.This study shows that increased cardiac levels of the human and murine BBLN protein promoted cardiac dysfunction and adverse cardiac remodeling in Tg-BBLN mice and in mice with chronic pressure overload-induced cardiac dysfunction.BBLN triggered cardiac damage in a protein dosage-dependent manner by interaction with CAMK2D and enhancement of the activating CAMK2D autophosphorylation on T287 in vitro and in vivo.Phosphorylated phospho-T287-CAMK2D exerts autonomous kinase activity, which is independent of its stimuli, calcium and calmodulin 23 .By CAMK2D activation, BBLN promoted CAMK2D-dependent adverse cardiac remodeling with cardiac dysfunction, fibrosis and necroptosis.Vice versa, BBLN-induced cardiac pathologies and necroptosis were retarded by RNAi-mediated CAMK2D protein downregulation.Furthermore, a BBLN mutant with defective CAMK2D interaction (BBLN-SxxA), which did not act as an activating CAMK2D substrate, was largely inert in vitro and did not stimulate cardiac dysfunction nor negative cardiac remodeling in vivo, in transgenic Tg-BBLN-SxxA mice.
In vitro experiments found that BBLN interacts with the CAMK2D kinase domain and acts as an activating CAMK2D substrate.Sequence alignment showed that the human BBLN amino acid sequence displayed major features, which are characteristic for CAMK2 kinase domain interaction and which are conserved in CAMK2-activating substrates 28,29 .Likewise, the interaction of BBLN with the CAMK2D kinase domain was reduced in the presence of the prototypical CAMK2-activating peptide of GRIN2B.In addition to CAMK2D activity enhancement, human BBLN interacts with intermediate filaments 42 .In this respect, human BBLN resembles the Caenorhabditis elegans ortholog of BBLN, which also interacts with intermediate filaments 44 .However, the major CAMK2-activating features are apparently missing in the amino acid sequence of the C. elegans ortholog of BBLN, with only 15.9% identity with the human BBLN amino acid sequence 43,44 .Therefore, the pathological phenotypes of the human BBLN protein and the C. elegans ortholog could be different.Article https://doi.org/10.1038/s44161-023-00351-6 We found that BBLN was upregulated in human heart specimens of TOF patients with cyanosis.Upregulation of the human BBLN protein could be triggered by hypoxia in TOF patients with cyanosis because the human BBLN gene is an EGR1-responsive gene, and EGR1 was found to be induced in cyanotic TOF patient hearts 15 .BBLN expression correlated with markers of adverse cardiac remodeling in right ventricular TOF patient heart specimens with cyanosis and in right ventricular heart tissue of Tg-BBLN mice.The increased BBLN levels could contribute to the cardiac deterioration of TOF patients because elevated cardiac BBLN levels caused cardiac remodeling and cardiac dysfunction in vivo, in Tg-BBLN mice subjected to pressure overload, in a protein dosage-dependent manner Taken together, this study identified BBLN as a pathologic factor in TOF patients with hypoxia and in mice overexpressing BBLN under cardiac pressure overload.The upregulated BBLN protein, at least in the mouse model, promotes adverse cardiac remodeling and predisposes to an increased heart failure risk.One of the limitations of the current study is that in vivo work was not done in a mouse model of TOF, but rather in a model of pressure overload as a proxy.As hypoxia is a major pathological feature of TOF patients worldwide, the identification of the elucidated BBLN-CAMK2D pathway may account for a substantial disease burden of many TOF patients worldwide 45 .In addition, BBLN could contribute to the cardiac deterioration of patients with other hypoxia-induced forms of heart disease.Consequently, future treatment modalities of hypoxic conditions in the heart could aim to target pathologic BBLN functions, for example, by RNAi-mediated downregulation of BBLN or by interference with the BBLN-enhanced CAMK2D kinase activation.

Ethical compliance
The research complies with all relevant ethical regulations.

Generation of transgenic mice and animal experiments
The following transgenic mouse lines were generated and used: FVB/N-Tg(MHCBBLN) Sjaa (Tg-1), FVB/N-Tg(MHCBBLN)2 Sjaa (Tg-2), FVB/N-Tg(MHCBBLN) This study used male mice and female mice, and investigated the phenotype of transgenic mice in comparison to age-and sex-matched nontransgenic mice at an age of 3-4 months and 8 months as indicated.Sample size of animal experiments was predetermined.
No randomization was performed.Experiments with transgenic mice were routinely performed with mice from at least three different breeder pairs.At the end of the observation period, mice were anesthetized by intraperitoneal (i.p.) injection of ketamine and medetomidine (75 mg kg −1 and 0.5 mg kg −1 body weight, respectively), and cardiac function parameters were measured in the parasternal long-axis view by an observer, who was blinded to the genotype by echocardiography in M-mode with a Vivid 7 ultrasound system using a M12L linear array transducer (GE Healthcare).Data evaluation was performed with the EchoPAC software version 3.0 (GE Healthcare), and the left ventricular cardiac EF was determined by the formula of Teichholz.For isolation of RNA and proteins, transcardial perfusion of mice was performed with phosphate-buffered saline (PBS), under terminal anesthesia (ketamine and xylazine, 200 mg kg −1 and 60 mg kg −1 body weight, respectively, i.p.).Hearts were immediately frozen in liquid nitrogen or processed for histology. Animal

Human TOF heart specimens
The study analyzed human heart specimens of the RVOT from pediatric patients during clinically indicated primary cardiac repair surgery for TOF.For RNA extraction, cardiac specimens (muscle bundles that are routinely resected from the RVOT as part of surgery) were recovered of 11 pediatric patients with sporadic TOF cardiac defects and cyanosis (repeated oxygen saturation measurements of ≤91.0% on room air, mean 85.3 ± 3.3%, and episodes of cyanotic spells; age 23.2 ± 5.0 months; four males and seven females).For RNA extraction, cardiac specimens were collected in RNAlater (Qiagen GmbH, no.1017980), immediately frozen and transferred to −80 °C for long-term storage.In addition, cardiac specimens were obtained from other 16 TOF patients with cyanosis (oxygen saturation ≤91.0%, mean 86.7 ± 3.5% and episodes of cyanotic spells; age 24.4 ± 5.7 months; seven males and nine females) and 16 TOF patients without cyanosis (oxygen saturation ≥92%, mean 94.3 ± 1.6% and no episodes of cyanotic spells; age 28.8 ± 6.8 months; eight males and eight females).These 32 cardiac specimens were cut into two pieces and used for histological analysis (collected in 10% formalin), and protein extraction (collected in radioimmunoprecipitation assay (RIPA) buffer supplemented with protease/phosphatase inhibitor cocktail, immediately frozen and transferred to −80 °C for longterm storage).
All pediatric patients had sporadic TOF cardiac defects without any other malformation and had no 22q11 deletion.Informed consent was obtained from all parents.There was no participant compensation.The study protocol was performed in compliance with all relevant ethical regulations and was approved by the ethical committee of the MRC, Ain Shams University Hospital, Cairo, Egypt (date of approval 18 October 2006).

Cardiac transcriptome profiling by NGS and whole-genome microarray gene expression analysis
For transcriptome analysis by NGS, right and left ventricular mouse heart specimens from 3-4-month-old male Tg-BBLN (Tg-1) mice, and age-and sex-matched nontransgenic FVB mice were dissected, pulverized under liquid nitrogen and total RNA was isolated by the RNeasy Midi kit (Qiagen GmbH, cat.no.75144).For transcriptome sequencing by NGS, RNA libraries were prepared using standard Illumina protocols in frame of the INVIEW Transcriptome Discover protocol (Eurofins Genomics Germany GmbH).The hearts of 8-month-old, male Tg-BBLN (Tg-1) mice and nontransgenic FVB controls were similarly processed Article https://doi.org/10.1038/s44161-023-00351-6 for NGS.Illumina paired-end read sequencing (2 × 150 bp) with guaranteed 30 million read pairs per sample was done by the Genome Sequencer Illumina HiSeq, in the sequence mode HiSeq 4000.RNA sequencing reads in FASTQ format were imported into the program CLC Genomics workbench 20 version 20.0.4 (Qiagen Bioinformatics) and mapped to the reference genome (Mouse GRCm39) in frame of the RNA Sequencing Workflow of CLC Genomics Workbench 20.NGS transcriptome profiling of right ventricular heart tissue of Tg-BBLN mice was validated by the detection of right ventricle-specific transcript changes in comparison to left ventricular tissue (Extended Data Fig. 4).Transcript selection is based on literature data of differential gene expression between right and left ventricles from hypoxic rats, heart failure rats and/or right ventricular heart failure patients [46][47][48] .
Whole-genome microarray gene expression profiling of 11 human TOF heart specimens was performed with frozen cardiac specimens from 11 pediatric TOF patients with cyanosis (four males and seven females; age 23.2 ± 5.0 months).The synthesis of complementary DNA, labeling and hybridization of fragmented cRNA (15 μg) with the gene chip (Affymetrix GeneChip Human Genome U133 Plus 2.0 Array) were performed according to the Affymetrix protocol (Affymetrix GeneChip Expression Analysis Technical Manual Rev. 5).Microarray gene expression profiling was similarly performed of hearts from 8-month-old, male Tg-BBLN (low) mice and compared with those of ageand sex-matched nontransgenic FVB mice using the Mouse Genome MG430 2.0 Array (Affymetrix).The signals were processed using Affymetrix GeneChip Operating Software (v.1.4;Affymetrix).To compare samples and experiments, the trimmed mean signal of each array was scaled to a target intensity of 300.

IB detection of proteins
For IB detection of proteins, hearts were pulverized under liquid nitrogen, and proteins were extracted with RIPA buffer (150 mM NaCl, 1% NP40, 0.5% sodium deoxycholate, 0.1% sodium dodecyl sulfate (SDS) and 25 mM Tris, pH 7.4) supplemented with a protease/phosphatase inhibitor cocktail.Frozen human heart specimens, which were collected in RIPA buffer (supplemented with protease/phosphatase inhibitors), were thawed and lysed by sonication on ice.Insoluble material was removed by centrifugation followed by delipidation of solubilized cardiac proteins by the addition of acetone/methanol (12:2) at a final concentration of 83% for 90 min at 4 °C.The precipitate was washed with ice-cold acetone and dissolved for 90 min at room temperature in 6 M urea-containing SDS-sample buffer (with 2% SDS, 0.1 M dithiothreitol (DTT) and supplemented with a protease/phosphatase inhibitor cocktail).After the addition of iodoacetamide (10 mM), samples were stored at −80 °C until further use.Proteins were separated by 6 M urea-containing SDS-polyacrylamide gel electrophoresis (PAGE) (7.5% for proteins ≥40 kDa, and 10% for proteins <40 kDa) under reducing conditions.For IB detection, proteins were transferred to polyvinylidene fluoride membranes.After a blocking step, incubation with affinity-purified antibodies or F(ab') 2 fragments of the respective antibodies (preabsorbed to mouse/human serum proteins) at a dilution of 1:2,000-1:4,000 in blocking buffer was performed, and unbound primary antibodies were removed by washing steps.Thereafter, incubation with F(ab') 2 fragments of affinity-purified peroxidase-coupled secondary antibodies, preabsorbed to mouse/human serum proteins (dilution 1:40,000), or peroxidase-coupled protein A was performed.After additional washing steps, bound antibodies were visualized by enhanced chemiluminescent detection (Amersham ECL Prime, Cytiva RPN2236; Amersham ECL Select, Cytiva RPN2235).

AP of CAMK2D and CAMK2D-interacting proteins
To identify CAMK2D-interacting proteins, CAMK2D was AP from human heart specimens of TOF patients.The CAMK2D protein co-enrichment study pooled cardiac specimens of ten TOF patients, which were solubilized for 30 min at 4 °C in solubilization buffer (1% sodium deoxycholate, 0.05% SDS and 0.05% Tween 20 in PBS, pH 7.4, supplemented with protease/phosphatase inhibitors).Insoluble material was removed by centrifugation, and the supernatant was diluted 1:5 with PBS (supplemented with protease/phosphatase inhibitors) and applied to the immunoaffinity matrix (affinity-purified anti-CAMK2D antibodies coupled to Affigel 10; 5 mg IgG coupled to 1 ml of Affigel 10, Bio-Rad Laboratories, cat. no.1536099).Polyclonal anti-CAMK2D antibodies were raised in rabbit against full-length recombinant His 6 -CAMK2D expressed in and purified from baculovirus-infected Sf9 insect cells.After an overnight incubation at 4 °C, unbound proteins were removed by washing steps with PBS (20 column volumes), and bound proteins were rapidly eluted with 0.25 M NH 4 OH and 10% dioxane, pH 11.The pH of the eluate was immediately adjusted to pH 7.4.Eluted proteins were concentrated by centrifugation through Amicon Ultra centrifugal filters (Amicon Ultra-4 Centrifugal Filter Unit, molecular weight cutoff 3 kDa, UFC800308), dissolved in 2× SDS-Laemmli sample buffer and separated by SDS-PAGE under reducing conditions.Immunoaffinity-enriched CAMK2D protein and co-enriched BBLN were identified by IB detection.Co-enriched BBLN was also identified by nano-LC-ESI-MS/MS analysis (see below).

Identification of BBLN by nano-LC-ESI-MS/MS analysis
After enrichment of CAMK2D by AP from cardiac specimens of TOF patients, co-enriched BBLN was identified by nano-LC-ESI-MS/MS analysis.Isolated proteins were concentrated, dissolved in 2×SDS-Laemmli sample buffer, and separated by SDS-PAGE under reducing conditions.After staining of the SDS-containing polyacrylamide gel by Coomassie brilliant blue, visualized protein bands were cut out and subjected to nano-LC-ESI-MS/MS analysis.The nano-LC-ESI-MS/MS analysis and protein identification were performed by Proteome Factory AG.Data have been deposited to the PRIDE Proteomics Identifications Database (dataset identifier PXD044695).

Immunohistology and immunofluorescence
Immunohistological detection of BBLN was performed after antigen retrieval on cardiac specimens of TOF patients, and on longitudinal cardiac sections of Tg-BBLN mice and nontransgenic FVB mice, with affinity-purified, polyclonal anti-BBLN antibodies.Bound antibodies were visualized by an enzyme substrate reaction (DAB Enhanced Liquid Substrate System, cat.no.D3939, Sigma-Aldrich).Myocardial calcium accumulation as an indicator of necroptosis/necrosis was determined on cardiac specimens by the von Kossa calcium staining method (Diagnostic BioSystems, cat.no.KT028).Collagen staining of heart specimens by picrosirius red was performed by the Picrosirius Red Stain Kit (ab150681, Abcam).Immunohistological sections were imaged with a DMI6000 microscope equipped with a DFC 420 camera, and an MZ125 stereomicroscope equipped with a DFC 295 camera (Leica Microsystems).Colocalization studies of BBLN with CAMK2D, and of BBLN with phospho-S345-MLKL were performed by immunofluorescence applying (1) rabbit polyclonal anti-BBLN antibodies and mouse monoclonal anti-CAMK2D antibody (WH0000817M2; Sigma-Aldrich), and (2) rabbit polyclonal anti-BBLN antibodies and mouse monoclonal anti-phospho-S345-MLKL antibody (MABC1158, Clone 7C6.1;EMD Millipore Corporation) (dilution 1:200).Secondary antibodies were labeled with Alexa Fluor 488 and Alexa Fluor 568, respectively (dilution 1:4,000).Nuclei were stained with 4,6-diamidino-2-phenylindole.Sections were imaged with a confocal laser scanning microscope (Leica TCS SPE, and Leica SP8 Falcon; Center for Microscopy and Image Analysis at the University of Zurich).

Statistical analyses and software
The presented data show biological replicates unless otherwise stated.All the nonhuman data of this study have been confirmed at least three times with consistent results.Applied statistical tests are indicated in the figure legends, and data are presented as mean ± s.d.Comparisons between two groups were made with the unpaired, two-tailed t-test, and for comparisons between more than two groups, ordinary one-way analysis of variance (ANOVA) with a post-test as indicated was used.Survival analyses were performed by Kaplan-Meier survival analysis with a log-rank (Mantel-Cox) test.Linear regression analysis was performed for linear dependence analysis between two variables, and the Pearson correlation (r) was determined.A p-value of <0.05 was considered as statistically significant.Only transcripts showing correlation with BBLN in human RVOT specimens of cyanotic TOF patients with a P value of <0.05 and a Pearson correlation (r) of ≥+0.6 or ≤−0.6 were included in the overrepresentation analysis.Statistical analyses were performed with R, and the linear regression analysis was performed with Excel 2019 (version 16.0).GraphPad Prism (version 9.3.1)was used for creation of graphs.NGS transcriptome data comparisons between two groups were performed by MeV, and the unpaired, two-tailed t-test (just alpha) 49 .The heat map was generated by Morpheus (https://software.broadinstitute.org/morpheus).The overrepresentation analysis was performed with g:GOSt of g:Profiler (version e107_eg54_p17_bf42210, database updated on 15.09.2022).P values were determined with Fisher's one-tailed test 50 .Multiple testing correction was performed with the G:SCS algorithm of g:GOSt 50

Fig. 1 |
Fig. 1 | Identification of the CAMK2D-interacting protein BBLN in cardiac specimens of TOF patients.a, Identification of BBLN in TOF patient heart specimens by nano-LC-ESI-MS/MS analysis of CAMK2D-co-enriched proteins in the 8-14 kDa range.b, AP of CAMK2D from cardiac specimens of TOF patients (AP: CAMK2D) followed by IB detection of enriched CAMK2D (IB: CAMK2D) and co-enriched BBLN (IB: BBLN).The control immunoaffinity matrix (AP: con) did not enrich CAMK2D nor BBLN.The experiment was repeated three times with similar results.c, Characteristics of TOF patients who were included in the study for the immunohistological determination of cardiac BBLN.Age between acyanotic and cyanotic TOF patients was comparable while oxygen saturation (sat.) was significantly different.d, Immunohistological detection of BBLN on

Fig. 2 |
Female Tg-BBLN mice developed a heart failure phenotype with an increased mortality.a, Decreased lifespan of female Tg-BBLN mice (Tg-1 and Tg-2), was detected by Kaplan-Meier survival analysis with log-rank (Mantel-Cox) test (n = 103 mice per group; p < 0.0001).b, Symptoms of heart failure with a significantly decreased left ventricular ejection fraction (EF) were detected by echocardiography in female Tg-BBLN mice (Tg-1) at an age of 3-4 months.c, Female Tg-BBLN mice (Tg-1) also developed cardiac hypertrophy as documented by an increased left ventricular internal diameter in diastole (LVIDd).Data (b,c) are mean ± s.d.(n = 6 female mice per group; age: 3-4 months; unpaired, two-tailed t-test; df=10; b, t = 12.56; p = 1.89718E-07; c, t = 4.113; p = 0.0021).Controls were age-matched, non-transgenic, female FVB mice.Extended Data Fig. 3 | Whole genome microarray gene expression profiling of tetralogy of Fallot (TOF) patient heart specimens.a, Overview of microarray hybridization quality control features of whole genome microarray gene expression data of cardiac specimens from 11 patients diagnosed with TOF cardiac pathology and cyanosis (7 females, 4 males).Total cardiac RNA was extracted, processed for whole genome microarray gene expression profiling, and fragmented cRNA was hybridized to the gene chip (Affymetrix GeneChip Human Genome U133 Plus 2.0 Array).Raw data were analyzed with the GeneChip Operating Software (GCOS) and scaled to a target value of 300.b, Characteristics of 11 pediatric patients with TOF cardiac defects, from whom cardiac specimens were recovered for whole genome microarray gene expression profiling.Data are mean ± s.d.; n = 11 (7 females, 4 males).Extended Data Fig. 4 | Validation of right ventricular NGS transcriptome profiling data by detection of heart failure-related differences between right and left ventricular heart specimens of Tg-BBLN mice.a-j, NGS transcriptome data of right and left ventricular heart specimens of Tg-BBLN mice (Tg-1) were validated by detection of differential gene expression between right and left ventricles.

Data Fig. 5 |
Cardiac transcript levels of Camk2d splice variants in right and left ventricles of Tg-BBLN and FVB mice.a,e Cardiac transcript levels of different Camk2d splice variants were determined by NGS in right (a) and left ventricles (e) of 3-4-month-old, male non-transgenic FVB mice and transgenic Tg-BBLN mice (Tg-1), and are presented as TPM (transcripts per million).b-d, f-h, Transcript levels of cardiac Camk2a, Camk2b and Camk2g were determined by NGS in right ventricular (b-d) and left ventricular heart specimens (f-h) of non-transgenic FVB mice and transgenic Tg-BBLN mice.Data are mean ± s.d.
. As in TOF patient heart specimens with cyanosis, January 2007).The study protocol analyzing human heart specimens from pediatric TOF patients was performed in compliance with all relevant ethical regulations and was approved by the ethical committee of the MRC, Ain Shams University Hospital, Cairo, Egypt (date of approval 18 October 2006).Informed consent was obtained from all parents.There was no participant compensation.
. Multiple sequence alignment was performed with the Clustal Omega (CLUSTAL O, version 1.2.4) tool from European Molecular Biology Laboratory's European Bioinformatics Institute 51 .Pairwise sequence alignment was performed with EMBOSS Needle.A search of the National Center for Biotechnology Information GEO dataset browser identified upregulation of BBLN-Bbln in GEO datasets GDS2008 (GSE2299, ref. 52), GDS5302 (GSE54372, ref. 53) and GDS3018 (GSE4286, ref. 54).