68Ga-FAPI PET visualize heart failure: from mechanism to clinic

Heart failure (HF) is a chronic progressive clinical syndrome associated with structural and/or functional heart abnormalities. Active fibroblasts and ventricular remodelling play an essential role in HF progression. 68Ga-labelled fibroblast activation protein (FAP) inhibitor (68Ga-FAPI) binds to FAP. This study aimed to examine the feasibility of using 68Ga-FAPI positron emission tomography (PET)/computed tomography (CT) to visualize changes in cardiac fibrosis and function over time in the HF setting. After establishing an isoproterenol (ISO)-induced HF rat model (14 consecutive days of intraperitoneal ISO injections), echocardiography and 68Ga-FAPI PET/CT were performed weekly in experimental and control groups. Rat hearts were examined weekly for biodistribution analysis; autoradiography; and haematoxylin and eosin, FAP immunofluorescence and Masson’s trichrome staining analysis. Rat blood was sampled weekly for enzyme-linked immunosorbent assay analysis of various plasma indicators. A preliminary clinical study was also performed in seven HF patients who underwent both 13N-amino (NH3) perfusion and 68Ga-FAPI cardiac PET imaging. In the animal experiments, myocardial 68Ga-FAPI uptake, expression of FAP and myocardial contractility peaked on day 7 after the initial ISO injection. Only slight fibrotic changes were observed on histopathological examination. 68Ga-FAPI uptake and ventricular wall motion decreased over time as cardiac fibrosis and degree of myocardial injury gradually increased. In the human HF patient study, 68Ga-FAPI PET imaging identified varying degrees of 68Ga-FAPI uptake in the myocardium that did not precisely match with 13N-NH3 myocardial perfusion. As HF progresses, 68Ga-FAPI uptake is high in the early stages and then gradually decreases. Although preliminary, our findings suggest that 68Ga-FAPI PET can be used to demonstrate active myocardial fibrosis. Active myocardial FAP expression is followed by myocardial remodelling and fibrosis. Detection of early active FAP expression may assist treatment decision making in HF patients. Clinical Trial Registration: NCT04982458


Introduction
Heart failure (HF) is a chronic and progressive clinical syndrome in which the heart structure or function is abnormal owing to multiple factors [1]. Worldwide prevalence of HF is 64.3 million people and rising [2]. Survival of HF patients decreases as the disease progresses [3]. Considering that HF-related mortality is similar to mortality rates related to prostate and colorectal cancer in men and breast and ovarian cancer in women [4], HF can be regarded as cardiovascular "malignancy". Understanding HF progression and related factors is imperative to ensure optimal and cost-effective treatment [5].
Fibrosis is observed in nearly every type of myocardial disease [6]. After heart injury, cardiac fibroblasts (CFs) begin to remodel the myocardium via deposition of extracellular matrix (ECM), which results in increased tissue stiffness and reduced compliance [7,8]. Excessive CFs is an important factor in HF progression [9]. Currently, endomyocardial biopsy and magnetic resonance imaging (MRI) can be used to evaluate CF density and pathology [10]. However, biopsy is invasive and MRI cannot detect pathological CFs in the early phase of disease. Fibroblast activation protein (FAP) is robustly expressed in activated CFs [11] and is a potential marker of HF progression.
Clinical therapy targeting cardiac fibrosis is limited [12]. Angiotensin-receptor-neprilysin inhibitors improve ventricular remodelling and delay fibrosis progression [13]. Pirfenidone [14] and engineered T cells are currently in clinical studies [15]. Early and precise identification of CFs before treatment might be important. Moreover, FAP quantitation may provide information regarding treatment response. 68 Ga-labelled FAP inhibitor ( 68 Ga-FAPI) is a novel agent that targets FAP and shows the activated fibroblast response with a high target-to-background ratio [16,17]. Previous studies have shown that 68 Ga-FAPI can visualize fibrosis in the lung and liver [18] and detect FAP expression in acute myocardial infarction [19]. However, its role in detecting CFs in HF has not been examined.
In this study, 68 Ga-FAPI positron emission tomography (PET)/computed tomography (CT) imaging was performed in an isoproterenol hydrochloride (ISO)-induced HF rat model to examine its feasibility for visualizing changes in CFs by comparing imaging changes with histopathological changes over time. As a preliminary clinical study, we also used 13 N-amino (NH 3 ) and 68 Ga-FAPI PET imaging to evaluate myocardial perfusion and fibrosis simultaneously in HF patients; our aim was to examine the suitability of using 68 Ga-FAPI PET to quantify cardiac FAP and visualize cardiac fibrosis in HF patients.

HF models
Animal experiments were approved and conducted under the guidance of the animal ethics committee of Tongji Medical College of Huazhong University of Science and Technology. Fifty Sprague-Dawley rats (weight, 190 ± 20 g) were divided into an ISO-induced HF group (n = 25) and control group (n = 25). ISO-induced HF group rats received ISO 5 mg/kg/day via intraperitoneal injection for 14 days [20]. Rats in the control group received intraperitoneal injections of the same volume of physiological saline over the same period.

Echocardiographic assessment of cardiac function
Routine transthoracic echocardiography was performed every week before and after ISO administration by two experienced operators (S He and J Wang) to verify the HF model and assess left ventricular structural and functional parameters. Rats were anaesthetized before echocardiography using 1.5 to 2% isoflurane in the supine position at room temperature. Echocardiography was performed using the Vivid E95 ultrasound system with a 12S probe at 9.0 MHz and EchoPAC workstation (GE Medical Systems, Milwaukee, WI, USA). M-mode images at the level of the left ventricle papillary muscles were acquired and guided by two-dimensional parasternal short-axis views. Left ventricular ejection fraction (LVEF), left ventricular fraction shortening (LVFS), systolic left ventricle internal dimension (LVIDs), diastolic left ventricle internal dimension (LVIDd), systolic interventricular septal thickness (IVSs), diastolic interventricular septal thickness (IVSd), systolic posterior wall thickness (LVPWs), diastolic posterior wall thickness (LVPWd), left ventricular end-diastolic volume (LVEDV), left ventricular end-systolic volume (LVESV), stroke volume (SV) and left ventricular mass (LVM) were measured.

Preparation of 68 Ga-FAPI and microPET/CT imaging
The methods used for 68 Ga-FAPI synthesis and labelling have been previously described [21]. Experimental and control group rats were injected with 68 Ga-FAPI (18.5 ± 0.7 MBq) via the tail vein on day 0 before the first ISO injection and days 7, 14, 21 and 28 after the first ISO injection. Micro-PET/CT (TransPET Discoverist 180, Raycan Technology, SZV, CHN) imaging was performed 45 min after 68 Ga-FAPI injection using an acquisition time of 15 min. MicroPET/ CT images were reconstructed with the ordered-subset expectation maximization three-dimensional/maximum a posteriori probability algorithm and then analysed using Inveon Research Workplace (Siemens, Munich, Germany).

Tissue biodistribution and autoradiography
On days 0, 7, 14, 21 and 28, experimental and control group rats (three from each group) were sacrificed 30 min after 68 Ga-FAPI injection. The hearts were then harvested and weighed. Radioactivity was quantified using the Wizard 2470-γ-counter (PerkinElmer, Waltham, MA, USA). The heart was then serially sectioned as shown in Fig. 1a (section thickness, 2 mm) and biodistribution measurements were performed. The biodistribution results are expressed as a percentage of the injected dose per gram of tissue (%ID/g). Next, phosphor screen autoradiography (ARG) was performed by exposing tissue slices on the phosphor screen for 15 min followed by scanning with the Cyclone Plus Storage Phosphor Imaging System (PerkinElmer).

Histology examination
To examine myocardial fibrosis, rat heart tissue specimens were collected on days 0, 7, 14, 21 and 28 for haematoxylin and eosin and Masson's trichrome staining. Each specimen was fixed using 4% paraformaldehyde and embedded in paraffin. FAP expression was identified in specimens using immunofluorescence (IF). Fixed tissue sections were treated in methanol at −20°C for 15 min in a serum-free blocking agent for 1 h at room temperature and then incubated with the primary antibody for 1 h. After incubation with the secondary antibody (FAP antibody, 1:200 dilution; Abcam, Cambridge, UK) for 40 to 60 min, the sections were mounted in medium with 4′,6-diamidino-2-phenylindole. Mean grey values of the FAP protein were determined using immunofluorescence. Collagen volume fraction (CVF) was calculated from Masson's trichrome-stained slides as collagen area/total observed area × 100%. Image J software (National Institutes of Health, Bethesda, MD, USA) was used for image analysis.

Enzyme-linked immunosorbent assay
Blood samples were collected at different times from the abdominal aorta in ethylenediaminetetraacetic acid-K2 vacuum tubes and centrifuged for 15 min at 3000 × g to obtain plasma. Concentrations of plasma C-reactive protein (CRP), interleukin 6 (IL-6), cardiac troponin I (cTn-I), tumour necrosis factor alpha (TNF-α), angiotensin II (Ang-II) and B-type natriuretic peptide (BNP) were measured as cardiacrelated indicators using a commercially available enzymelinked immunoassay kit (Ruixi Biotech, Inc., ZLSN, CHN) according to the manufacturer's protocol.

Patients
The patient component of this study was approved by the clinical research ethics committee of Union Hospital (no. 20210617-01) and registered at the Clinical Trail (NCT04982458). Inclusion criteria were as follows: (1) symptoms and signs of HF, (2) clinical evidence of cardiac abnormality, (3) age ≥18 years and (4) complete clinical data. We excluded patients with acute coronary syndrome, acute systemic disease or infection, uncontrolled metabolic disease and severe hepatic or renal dysfunction. Pregnant and lactating women were also excluded. All participants provided written informed consent. All patients underwent both 13 N-NH 3 myocardial perfusion imaging and 68 Ga-FAPI PET on the same day. Retrospective cardiac 68 Ga-FAPI uptake data from 20 subjects without cardiac disease were used as a control group.

Ga-FAPI and 13 N-NH 3 imaging protocol and image analysis
For 13 N-NH 3 perfusion imaging, PET (Discovery VCT; GE Healthcare) was performed immediately after intravenous injection of 13 N-NH 3 (370-740 MBq; acquisition time, 10 min). Two hours later, 68 Ga-FAPI (1.8-2.2 MBq/kg) was injected intravenously and imaging was performed for 20 min at 45-min time points as previously described [22]. Attenuation-corrected PET images were reconstructed using the ordered-subset expectation maximization iterative method. The gated data were reconstructed using volume image protocol replay. 13 N-NH 3 perfusion and 68 Ga-FAPI imaging were further processed using Emory Cardiac Toolbox in the Xeleris Workstation 2.0 (GE Healthcare). The following PET indices were acquired using the Advantage Workstation AW4.6 (GE Healthcare): 68 Ga-FAPI maximum standardized uptake value (SUV max ), 68 Ga-FAPI mean standardized uptake value (SUV mean ) and SUV max of myocardium/cardiocoelomic SUV mean (normalized SUV max ).

Statistical analysis
Data are expressed as means with standard deviation. Statistical analyses were performed using Prism 8 software (GraphPad, San Diego, CA, USA) and SPSS software version 13.0 (IBM Corp., Armonk, NY, USA). Comparisons were performed using analysis of variance and the Student's t test as appropriate. Correlation was evaluated using Pearson's method. P < 0.05 was considered significant.

Ga-FAPI phosphor screen ARG and histopathology
The hearts were dissected and sliced into consecutive sections as shown in Fig. 1a. 68 Ga-FAPI uptake in HF group rats was increased on days 7 and 14 and then gradually decreased on days 21 and 28 (Fig. 1b). Furthermore, cardiac radioactivity in the coronal plane significantly decreased from the apical cordis to the basal cordis. Haematoxylin and eosin and Masson's stain examinations showed that fibrosis increased from day 0 to day 28 in HF group rats. FAP IF showed that FAP expression reached its peak on day 7, began to decrease on day 21 and was undetectable on day 28 (Fig. 1c).

Ga-FAPI imaging
No obvious 68 Ga-FAPI uptake was observed in the myocardium on day 0 in the HF and control groups (Fig. 2a).
In the HF group, extensive myocardial uptake was observed after ISO injection: peak uptake occurred on day 7 and then decreased from day 14 to day 28. Echocardiography showed that as HF progressed, ventricular systolic function worsened (Fig. 2b). M-mode images showed peak ventricular wall thickening on day 7 in conjunction with increased amplitude of ventricular wall motion and enhanced myocardial contraction. After day 7, the heart chambers gradually expanded, the thickness of the ventricular wall gradually decreased, the amplitude of ventricular wall motion gradually decreased, and myocardial contraction gradually weakened.

Cardiac dysfunction in the HF group
Echocardiographic parameter data over time in the HF group are shown in Fig. 3. LVEF, LVFS, LVPWd, LVPWs, IVSd, IVSs and LVM significantly increased on day 7 and then gradually decreased, implying a decrease in ventricular systolic function after excessive compensation. However, LVEDV, LVESV, SV, LVIDd, LVIDs and cardiac volume significantly decreased on day 7 and then increased over time.

Cardiac enzyme-linked immunosorbent assay in the HF group
Plasma concentrations of BNP, cTn-I, CRP, Ang-II, IL-6 and TNF-α in the HF group are shown in Fig. 4. BNP concentration was significantly higher on day 14 than day 0. Importantly, concentrations of cTn-I, CRP, Ang-II and IL-6 levels were significantly higher on days 14, 21 and 28 than day 0. However, TNF-α concentration was significantly higher on days 21 and 28 than day 0.

In vitro biodistribution and pathological section statistical analysis of the HF group
On day 0 before ISO injection, cardiac uptake of 68 Ga-FAPI did not significantly differ among the myocardial layers. The heart-to-muscle (H/M) ratio of 68 Ga-FAPI was highest on day 7 and then gradually decreased (Fig. 5a). The site of highest concentration was the Heart 5 (apex cordis) section, where H/M ratio and uptake values were 19.97 ± 0.33 and 0.61 ± 0.07 %ID/g on day 7, respectively (Supplementary Tables 1  and 2).

Ga-FAPI and 13 N-NH 3 imaging results in HF patients
Seven HF patients (5 men and 2 women) who met criteria were included for analysis. Mean age was 58.14 ± 16.25 years (Table 1). 68 Ga-FAPI normalized SUV max in the heart was significantly higher in the HF patients than the control patients (5.43 ± 0.99 vs. 1.24 ± 0.25; P < 0.001). Cardiac uptake varied in HF patients as shown in Fig. 6a.
Myocardial perfusion and 68 Ga-FAPI uptake were not consistent with each other in HF patients. Two representative patients are shown in Fig. 6b and c to demonstrate the characteristics of 68 Ga-FAPI and 13 N-NH 3 imaging. Case 1 was a 72-year-old female with a history of congestive dilated cardiomyopathy who complained of chest discomfort and shortness of breath for 1 month. On echocardiography, left ventricular wall motion was reduced, the left heart was enlarged, and LVEF was 25%. 13 N-NH 3 PET showed myocardial perfusion was reduced to varying degrees in the apex, middle segment and basal segments of the left ventricle. Avid accumulation of 68 Ga-FAPI were observed in the left ventricle, especially the anterior and inferior walls (normalized SUV max , 9.0). Case 2 was a 46-year-old male with a history of coronary heart disease who complained of slowly progressive shortness of breath after activity for 9 months. Echocardiography showed enlargement of the left heart and an LVEF of 24%. On 13 N-NH 3 PET, a large uptake defect was visualized in the left ventricular septum. However, 68 Ga-FAPI PET showed obvious uptake in the anterior and inferior wall (normalized SUV max , 8.89). Combined analysis showed that 13 N-NH 3 perfusion was inconsistent with 68 Ga-FAPI uptake, and partial supplemental defect imaging presented. 68 Ga-FAPI uptake was widely distributed throughout the ventricle to varying degrees and was remarkable in the anterior and inferior walls. We hypothesize that this is explained by active expression of FAP in the ventricular remodelling phase of HF. In case 1, 13 N-NH 3 imaging showed a slightly uneven reduction in perfusion. In case 2, defects of 13 N-NH 3 and 68 Ga-FAPI PET uptake were shown in part of Fig. 5 Quantitative analysis of HF models. a Heart-to-muscle (H/M) uptake ratios at different times (0, 7, 14, 21 and 28 d) and sections (Heart 1-5 and total). b Linear plots of collagen volume fraction (CVF) and mean grey value over time (red, CVF; blue, mean grey value). c Correlation analysis between H/M ratios of Heart 5 and mean grey value (green dots, variables). The cross section of heart section is shown in Fig. 1a. * refers to the ratio of H/M in the same section with 0 d. * refers to P < 0.05; ** refers to P < 0.01; *** refers to P < 0.001; **** refers to P < 0.0001. + refers to different groups compared with 0 d group. + , P < 0.05; ++ , P < 0.01; +++ , P < 0.001; ++++ , P < 0.0001 the left ventricular septum, which implied old myocardial infarction. However, 13 N-NH 3 perfusion was deficient in parts of the apex and proximal apex, which showed moderate 68 Ga-FAPI uptake, suggesting the area was actually severely ischemic with active FAP expression.

Discussion
In this study, 68 Ga-FAPI PET was used to visualize CFs and monitor HF progression in rat model of ISO-induced HF. 68 Ga-FAPI accumulated in the heart and reached a peak after the first injection of ISO, then decreased as HF initiated and progressed. 68 Ga-FAPI uptake findings correlated well with those of the myocardial Masson's and FAP IF staining. The 13 N-NH 3 perfusion and 68 Ga-FAPI cardiac PET imaging in human HF patient component of the study demonstrated varying degrees of myocardial 68 Ga-FAPI uptake that was not consistent with the myocardial perfusion findings. These preliminary results suggest that 68 Ga-FAPI PET may visualize fibre activation in the early stages of fibrosis in HF. Therefore, 68 Ga-FAPI PET may enable identification of HF patients in the early active stages of fibrosis, which might assist with treatment decision making.
Cardiac remodelling is currently considered one of the most important clinical determinants of HF progression [23]. Remodelling involves typical molecular, cellular and extracellular changes that manifest clinically as variations in heart size, shape and function [24]. The pathophysiology and underlying cellular processes that drive cardiac remodelling are not caused by loss of cardiomyocytes or focal myocardial injury, but rather by pervasive and global fibroblast proliferation and development of fibrosis [25]. Therefore, detecting these processes and quantifying the degree of fibrosis should enable better understanding of cardiac remodelling and provide a scientific basis for early HF prevention and treatment. FAP is highly expressed in activated fibroblasts and reflects the fibrosis process in numerous diseases [26].
Cardiac uptake of 68 Ga-FAPI peaked on day 7 and then gradually decreased, which was confirmed on ARG and immunoprotein staining. This demonstrated that 68 Ga-FAPI could visualize dynamic changes in FAP expression [27,28]. Cardiac function reached its maximum on day 7 and then decreased, which mirrored the 68 Ga-FAPI uptake findings. In contrast to the FAP expression changes, histological validation of CVF revealed that myocardial fibrosis started on day 14 and gradually progressed over time. Similarly, ventricular load, myocardial cell necrosis and inflammation of the cardiovascular system gradually worsened after day 14 as reflected by plasma indicators. Therefore, a large area of avid FAP expression may trigger extensive myocardial fibrosis and result in HF. Active FAP expression may also serve as an indicator for early anti-ventricular remodelling treatment. Thus, 68 Ga-FAPI might be a useful tool to visualize early HF progression and assess the efficacy of antifibrotic therapy. 68 Ga-FAPI and 13 N-NH 3 imaging in HF patients confirmed FAP expression and suggested that 68 Ga-FAPI is sensitive for detecting active fibrosis. These imaging modalities appear to provide indirect evidence of CF numbers and activation. In addition, 68 Ga-FAPI uptake was inconsistent with 13 N-NH 3 perfusion. Using both allows visualization of myocardial perfusion and CF activation. Old myocardial infarction demonstrates reduced uptake in both 68 Ga-FAPI and 13 N-NH 3 imaging. Active FAP expression and 13 N-NH 3 perfusion defect in an area may indicate severe ischemia with ventricular remodelling and pro-fibrotic activity. Our findings also support the theory that FAP overexpression has a role in both fibrosis of an ischemic region as well as reactive fibrosis in unaffected myocardium [29,30].
This study has several limitations. First, the number of HF patients was small and HF was not clinically staged. Further studies are warranted to examine the characteristics of 68 Ga-FAPI uptake in different stages of HF. Second, pathologic confirmation has not been performed in patients to exclude non-specific 68 Ga-FAPI uptake in myocardium, which is worth noting in the clinic [31]. Besides, the causes of HF in the examined patients were heterogeneous and different from the ISO-induced HF used in the rat model. The value of such imaging indicator should be confirmed in larger prospective studies. More patients should be included to better understand the pathological mechanisms of this finding.  13 N-amino ( 13 N-NH 3 ) images in patients without cardiac disease and with HF. Axial 68 Ga-FAPI PET/CT imaging of heart in control subjects (without cardiac disease) and HF patients (case 1, 2 and 7) (a). 13 N-NH 3 perfusion and 68 Ga-FAPI PET Bullseye plot and short-axis serial images in patient cases 1 (b) and 2 (c), respectively

Conclusion
As HF progresses, 68 Ga-FAPI uptake is high in the early stages and then gradually decreases to almost indiscernible levels. Active FAP expression prompts myocardial remodelling and fibrosis. Early active FAP expression may serve as an indicator for anti-ventricular remodelling treatment. 68 Ga-FAPI PET appears to be able to identify active myocardial fibrosis, which has treatment decision making significance.