Triglyceride (TG) and orphan diseases

TG is a major energy source for mammals. In normal condition, TG is either received via the diet, or synthesized endogenously and stored in adipose tissues. When required, TG is hydrolyzed by various enzymes called lipases and releases long-chain fatty acid (LCFA), which is delivered to non-adipose tissues for the production of ATP. It has been known that the ectopic TG deposition in non-adipose tissues causes some orphan diseases. In 1953, Jordans reported two brothers with phenotype of skeletal myopathy and vacuolar formation of peripheral leukocytes, called Jordans’ anomaly [1]. Fifty years later, Fischer et al. found that this phenotype is associated with mutations in PNPLA2 [2] encoding adipose TG lipase (ATGL) [3, 4], an essential molecule located in cytoplasmic lipid droplets for the intracellular TG hydrolysis [5, 6], and designated this phenotype as neutral lipid storage disease with myopathy (NLSD-M). Clinical manifestations of NLSD-M appeared variable from mild to severe symptoms [7,8,9,10,11,12,13], which could be at least partially explained by function of mutated ATGL proteins [14]. Another phenotype of NLSD involving the skin was reported as NLSD with ichthyosis (NLSD-I) by Chanarin and Dorfman in the 1970s [15,16,17]. The genetic cause of NLSD-I was found to be mutations in ABHD5 encoding CGI-58, a co-enzyme of ATGL [18]. Using skin fibroblasts and iPS cells from patients with NLSDs, unique intracellular metabolism of TG has been extensively analyzed. These cell-biological experiments showed that cytoplasmic lipid droplets are dynamic cellular organelles interacting with ATGL, CGI-58, and other proteins, and could be a therapeutic target [19,20,21,22,23].

Discovery of TG-deposit cardiomyovasculopathy (TGCV) with PNPLA2 (ATGL) mutation

Since the early 1980s, patients with Jordans’ anomaly and severe heart failure (HF), though very rare, had been reported in Japan [24]. In the early 2000s, our institution started to take care of two patients with severe HF and vacuolar formation in peripheral leukocytes. HF was progressive and intractable, and a couple of years later, they became candidates for cardiac transplantation (CTx). Preoperative examination of their hearts exhibited dilated cardiomyopathy-like morphology in chest X-ray and ultrasonography; however, endomyocardial biopsy specimens showed neutral lipid deposition in cardiomyocytes [25]. When they underwent CTxs, pathological and biochemical analyses of their explanted hearts were performed, demonstrating that their coronary arteries showed unusual coronary atherosclerosis with TG deposition in endothelial and smooth muscle cells (SMCs). We named this novel phenotype as TGCV [26,27,28]. These patients were identified as homozygous for genetic mutations in PNPLA2 encoding ATGL, which is also known to be responsible for NLSD-M as described above [2].

Postmortem analyses revealed undiagnosed individuals with TGCV

Retrospective postmortem analyses of autopsied cases identified individuals with TGCV phenotype who had TG deposit in both myocardium and coronary arteries, as presented in Fig. 1. A 38-year-old man suddenly died irrespective of intensive treatment for coronary artery disease (CAD) and HF. His heart was heavy in weight and hypertrophied with multiple myocardial fibrous scars. Coronary arteries showed diffuse and concentric stenosis in multi-vessels. Biochemical analyses and imaging mass spectrometry showed TG deposition in both myocardium and coronary arteries [29, 30]. TG-deposit SMCs were observed in his renal and mesenteric arteries as well (data not shown). These data mimic genetic ATGL deficiency; however, the immunoreactive mass of ATGL was detected, and the genetic test using genomic DNA extracted from stored specimens showed no mutation in all exons and exon/intron boundaries of PNPLA2 gene (data not shown). In addition, pathological records showed that he did not have skeletal myopathy.

Fig. 1
figure 1

Pathological analysis of the autopsied heart of a 38-year-old man with TGCV phenotype without PNPLA2 mutation. Panel a The transverse section of the autopsied heart stained with Masson’s trichrome showed circumferential patchy fibrosis of the left ventricular wall. The letters A, L, R, and P denotes anterior, left, right, and posterior, respectively. Panel b Lipid droplets (LDs) stained with oil red O in the cytoplasm of cardiomyocytes. Panel c Immunostaining for ATGL (Cell Signaling, Danvers, MA). Cardiomyocytes showed positive reactivity for ATGL. Panel d Coronary arteries with diffuse concentric-type stenosis. Panel e The transverse section of coronary artery was stained with Masson’s trichrome. Coronary artery revealed intimal thickening and fibroatheromatous lesions. Panel f Double-staining of Sudan black B and α-smooth muscle actin (Dako, Tokyo, Japan). Smooth muscle cells (brown color) with lipid droplet (blue color) distributed diffusely in the media and intima (arrows in Panel f). Asterisk represents the vascular lumen. Panel g TG (m/z 879.7) was identified as green and blue colors depending on the intensity. TG signals were diffusely detected in the arterial wall by imaging mass spectrometry. Green color denotes relatively higher intensity of TG than blue color. Myocardial and coronary TG contents (3.64 and 19.44 mg/g of tissue, respectively) were higher in this patient, compared with each of control group (1.4 ± 1.0, and 6.2 ± 4.8 mg/g of tissue, respectively). The detailed clinical profile of this patient is reported as Case 10 in the reference [29].

Scale bars: 1 cm in Panel a, 20 μm in Panel b and c, 5 mm in Panel d, 1 mm in Panel e, 20 μm in Panel f, 200 μm in Panel g

Development of diagnostic methods for TGCV

The above postmortem studies suggested that it is difficult to diagnose TGCV, and many undiagnosed patients should have died, which motivated us to develop diagnostic tools and methods for TGCV. We reported that myocardial scintigraphy with iodine-123-β-methyl iodophenyl-pentadecanoic acid (BMIPP) [31, 32], a radioactive analogue of LCFA, was useful in detecting abnormal LCFA metabolism in patients with TGCV [33, 34]. In addition, we reported the use of automated hematology analyzers to detect Jordans’ anomaly in patients with PNPLA2 mutation [35,36,37]. Recently, we developed CT-based TG imaging to detect myocardial and coronary TG deposition [34, 38] and selective immunoinactivation assay to measure functional ATGL activities using peripheral leukocytes [39].

Nomenclature, definition, and classification of TGCV

It is well known that disease nomenclature is made not only by their genotypes, but also by their phenotypes in many diseases and by discoverer’s names in some diseases. The nomenclature of TGCV was made by its phenotype that TG accumulated in both myocardium and coronary arteries, resulting from abnormal intracellular metabolism of TG and LCFA (Fig. 2) [26,27,28]. ATGL is a known enzyme involved in the phenotypic expression of TGCV. The Japan TGCV study group provided the diagnostic criteria for TGCV, in which TGCV with and without PNPLA2 mutations was designated as primary TGCV (P-TGCV) and idiopathic TGCV (I-TGCV), respectively [40,41,42].

Fig. 2
figure 2

Schematic presentation of the disease concept for TGCV

Pathophysiology of TGCV

The pathophysiological schema of TGCV is shown in Fig. 3. In normal condition (left panel, Fig. 3), LCFAs are taken up through transporters and receptors such as CD36. Some are transported to the mitochondria for β-oxidation, and the remaining LCFAs are utilized as a source of TG and rapidly hydrolyzed by intracellular lipases such as ATGL. In TGCV (right panel, Fig. 3), LCFAs are taken up and used to synthesize TG that cannot be hydrolyzed due to ATGL insufficiency, leading to energy failure and lipotoxicity with massive TG accumulation [28, 43]. It is emphasized that TG-deposit atherosclerosis is an important characteristic of TGCV [44] and distinct from usual cholesterol-deposit atherosclerosis, because the former showed diffuse and concentric narrowing formed by TG-deposit SMCs, whereas the latter showed discrete and eccentric stenosis initiated by the response to injury in the endothelium and accumulation of cholesterol-laden macrophages [45] (Fig. 4). We reported that TG-deposit SMCs and endothelial cells had pro-inflammatory and vulnerable phenotype in vitro [46, 47].

Fig. 3
figure 3

A pathophysiological model for TGCV. Genetic and acquired ATGL deficiency and other causes result in abnormal intracellular metabolism of TG and LCFA, leading to cardiomyocyte steatosis and TG-deposit SMCs. In normal condition (left panel), LCFA is taken up through LCFA transporters and receptors such as CD36 and some of them are transported to mitochondria for β-oxidation and the remaining LCFAs are utilized as a source for TG and rapidly hydrolyzed by intracellular lipases such as ATGL. In TGCV (right panel), LCFAs are taken up and used for the synthesis of TG which can not be hydrolyzed, leading to massive TG accumulation

Fig. 4
figure 4

Schemes for cholesterol- (Left) and TG-deposit atherosclerosis (Right). In cholesterol-deposit atherosclerosis, cholesterol (green) accumulates in macrophages, leading to eccentric stenosis. In TG-deposit atherosclerosis, TG (red) accumulates in SMCs, leading to concentric stenosis, which is a major feature of TGCV

A clinical case presentation of I-TGCV

A 58-year-old woman was referred to our hospital due to sudden chest tightness with ST-segment elevation in the electrocardiogram, followed by cardiopulmonary arrest. Under the diagnosis of acute myocardial infarction, she underwent coronary artery bypass grafting (CABG). Past history included type 2 diabetes mellitus requiring insulin treatment and hemodialysis. Cytoplasmic vacuoles in her peripheral polymorphonuclear leukocytes were observed less frequently (< 10% of neutrophils), compared with that in genetic ATGL deficiency (panel A in Fig. 5). ATGL activity in peripheral leukocytes was very low, comparable to that of genetic ATGL deficiency, as shown in Table 1. Myocardial washout rate (WOR) of BMIPP was defective in scintigraphy (panel B in Fig. 5). Pathological analyses of endomyocardial biopsy specimens demonstrated numerous vacuoles filled with stained lipid but positive reactivity for ATGL in cardiomyocytes and adipocytes (right, panel C in Fig. 5). Coronary CT angiogram showed diffuse narrowing coronary arteries, and in TG imaging [25], outside-in involvement of diffuse and abundant lipid components expressed as low CT numbers was seen within the wall in a peninsular pattern (arrows in panel D in Fig. 5). Her laboratory data and imaging tests were similar to those observed in TGCV with genetic ATGL deficiency, except for the conserved expression of ATGL protein in the myocardium. However, it is noted that the case was clinically distinct from genetic ATGL deficiency because there was no skeletal myopathy and no elevation of MM type creatine kinase. Genetic tests showed no mutations or substitutions in any of the exons or intron/exon boundaries of genes encoding ATGL, 1-acylglycerol-3-phosphate O-acyltransferase, hormone-sensitive lipase, or GOS2 (data not shown).

Fig. 5
figure 5

Laboratory and imaging examinations for TGCV. a Representative images of May-Giemsa staining of blood smears were shown from patients with P-TGCV and I-TGCV. b Bull’s eye images for BMIPP scintigrams from patients with P-TGCV and I-TGCV. The first scan was performed 20 min post-injection to determine early BMIPP uptake, and the second scan was performed 200 min later to study delayed uptake using myocardial SPECT after patients were injected with 123I-BMIPP. WOR was calculated with the Hear Risk View-S (HRSV) software as the difference between early and delayed images (reference value, 19.4 ± 3.2%). c A patient with P-TGCV showed numerous vacuoles (Panel a, H&E) in cardiomyocytes that stained positively for oil red O (ORO) (inset in Panel b). Furthermore, no positive reactivity for ATGL observed in any of cell types (Panel b, ATGL). Cardiomyocytes of patients with I-TGCV showed numerous vacuoles (Panel c, HE) filled with stained lipid (inset in Panel d, ORO), whereas positive reactivity for ATGL observed not only in adipocytes but also in cardiomyocytes (arrows in Panel d, ATGL). Scale bars: Panels a-d, 30 μm. d Coronary CT angiograms (CTA) from patients with P-TGCV and I-TGCV are shown. Bars in CTA correspond to Panels a-d, which are short axial sections of the left anterior descending coronary artery. The segmentation of the coronary artery lumen and wall was done using a workstation (Ziostation 2, Ziosoft, Japan). Constitutive components were classified into 4 colors with the original analysis software as follows. Colors indicate the CT number (yellow, − 25–0; orange, 0–40; green, 40–215; red, 215–700 Hounsfield unit [HU] (M@XNET, Tokyo, Japan) in Panels a-d. Yellow or orange areas indicate lipid components, red shows blood, and green shows the arterial wall without calcification or lipids. Black arrows in Panels a, b, and c indicate outside-in protrusion, which is the characteristics for TGCV

Table 1 Patients’ characteristics of Primary and Idiopathic TGCV

Clinical characteristics of P- and I-TGCV

Table 1 shows the clinical characteristics of 7 and 18 patients with P- and I-TGCV, respectively, registered to the international registry for NLSD and TGCV between February 2014 and March 2018 in Japan. Both TGCV types were adult onset with chest pain at rest or dyspnea and palpitation. Most patients with either types of TGCV developed severe HF or CAD with diffuse narrowing multivessel lesions or both. Myocardial metabolism of LCFA, detected by WOR of BMIPP and ATGL activities in peripheral leukocytes, was reduced in both TGCV types. Most patients with P-TGCV developed intractable and critical HF, as reported recently [26, 48, 49]. Two of them underwent CTx [26, 48]. Many patients with I-TGCV required percutaneous coronary intervention and CABG. As comorbidity, neither type of TGCV had skin lesions, which suggests that TGCV is not associated with NLSD-I. All patients with P-TGCV had skeletal myopathy, whereas none of those with I-TGCV did. Five of 7 and 3 of 18 registered patients with P- and I-TGCV, respectively, died.

Differential diagnosis of TGCV

Myocardial disorders such as dilated cardiomyopathy, hypertrophic cardiomyopathy, arrhythmogenic right ventricular cardiomyopathy, mitochondrial cardiomyopathy, alcoholic heart disease, and metabolic myocardial disorders (e.g., Fabry disease, Pompe disease, cholesteryl ester storage disease) need to be differentiated from TGCV [41, 42].

Furthermore, known diabetic and metabolic heart diseases need to be differentiated from TGCV. One is diabetic cardiomyopathy, which was originally defined as cardiomyopathy without significant stenosis in epicardial coronary arteries [50]. Another concept is epicardial fat accumulation, which is the oeverdeposition of TG in physiological tissues. TGCV is distinct from these two entities because TGCV is characterized by the ectopic deposition of TG in the cardiomyocytes and SMCs with apparent involvement of epicardial coronary arteries, as shown in Figs. 1 and 5.

Academia-initiated development of specific treatment for TGCV

We found that the chow with tricaprin, TG form of capric acid, improved LCFA metabolism, lipid deposition, cardiac function, and life span in ATGL-targeted mice [4], raising a therapeutic hypothesis that capric acid may be an alternative energy source and reduce TG deposition and lipotoxicity in TGCV [51]. Based upon these data, the Osaka University Hospital manufactured GMP-graded capsules containing the active gradients called CNT-01. We developed the assay to measure plasma capric acid levels [52, 53]. After finishing toxicity tests using rats and dogs required, we are finally conducting investigator-initiated clinical trials.

Comparison between NLSD-I, NLSD-M, and TGCV

As mentioned above, the nomenclature of TGCV was made by its phenotype that TG accumulated in both myocardium and coronary arteries, resulting from abnormal intracellular metabolism of TG and LCFA (Figs. 2 and 3). As described in the first paragraph in this letter, there have been known related disorders; NLSD-M and NLSD-I. Figure 6 shows the comparison of phenotype and genotype between TGCV and NLSDs. NLSD-M and NLSD-I are caused by mutations in PNPLA2 and ABHD5, mainly involved in the skeletal muscle and skin, respectively. Genotype of P-TGCV is known to be PNPLA2 mutation which is responsible for NLSD-M as well.

Fig. 6
figure 6

Relationship between TGCV and NLSDs. Comparison of phenotype and genotype between NLSD-I, NLSD-M and TGCV

Issues to be resolved

The following points are important focus for future researches:

  1. 1.

    Possible clinical continuum between P-TGCV and NLSD-M

    As mentioned above, both P-TGCV and NLSD-M is caused by genetic ATGL deficiency. It would be of interest to know whether patients with NLSD-M have TG-deposit atherosclerosis, which is the important feature for P-TGCV.

  2. 2.

    Etiologies of I-TGCV and its prevalence in countries other than Japan

    As shown in Table 1, 13 out of 18 patients with I-TGCV had family history of cardiovascular disease, suggesting that any genetic factors might be involved in the pathogenesis of I-TGCV. The mechanism underlying downregulation of ATGL activities of I-TGCV and possible involvement of other lipases and related enzymes is of significance to elucidate. In order to elucidate these issues, the development of screening methods for the diagnosis of I-TGCV is under way in our laboratory.

Conclusions

TGCV is a severe cardiovascular disorder named by its phenotype of cardiomyovascular TG deposition, of which etiologies seem heterogeneous.

Methods

  1. 1.

    Pathological, laboratory, and clinical imaging

    Standard procedures were performed as described (please see legends of Figs. 1 and 5).

  2. 2.

    International registry for NLSD/TGCV

    On the World Rare Disease Day 2014, we launched the international registry for neutral lipid storage diseases, TG-deposit cardiomyovasculopathy, and related disorders (Clinical Trial gov. NCT02830763).The present patients with TGCV were registered according to the study protocol after obtaining written consent. The protocol was approved by the Osaka University Hospital Ethical Committee (approval no. 13204).