Identification of Genes Related to 68Ga-DOTATATE Organ Distribution in Homo Sapiens


 Purpose 68Ga-DOTATATE is a somatostatin analogue that has been used for imaging neuroendocrine tumours. However, there is nonspecific uptake in some organs, and the reasons for that are not clear. The aim of the study is to outline the dynamic distribution pattern of 68Ga-DOTATATE in human body, and identify the genes responsible for 68Ga-DOTATATE uptake by bioinformatics analysis.Methods 68Ga-DOTATATE PET/CT was performed in 32 patients, and dynamic total-body PET scanning was performed with uEXPLORER. The gene expression datasets of human organs were downloaded from the Human Protein Atlas. WGCNA analysis was performed to screen the potential genes related to 68Ga-DOTATATE. BindingDB, SEA and SwissTargetPrediction databases were used to predict the potential binding proteins of DOTATATE based on molecular structure. Results Dynamic total-body PET scanning showed that 68Ga-DOTATATE uptake was not consistent with expression of SSTR2 in human organs. WGCNA analysis revealed 800 genes whose expression level was positively correlated to 68Ga-DOTATATE uptake. According to the molecular structure of DOTATATE, 135 proteins with potential binding ability to DOTATATE were screened by analysis based on drug database. Based on the results of the above two parts, 8 proteins were finally obtained, respectively AVPR1A, EPHA8, EPHB4, OPRL1, SSTR2, SSTR5, ST3GAL1 and NPY1R, which had potential binding possibility with DOTATATE, and their expression was positively correlated with 68Ga-DOTATATE uptake.Conclusion WGCNA analysis was combined with the drug database to obtain new potential binding targets of DOTATATE. The bioinformatics approach developed in this study could potentially be used to discover new potential binding targets for molecular imaging agents.


Introduction
Neuroendocrine tumors are tumors of neuroendocrine cells, usually associated with high expression of somatostatin receptor subtype 2 (SSTR2) [1]. 68 Ga-DOTA-D Phe 1 ,Tyr 3 -octreotate ( 68 Ga-DOTATATE), a somatostatin analogue with high a nity for SSTR2, has been used in the imaging of neuroendocrine tumors [2]. Once SSTR2 expression is con rmed by 68 Ga-DOTATATE based imaging, the same SSTR2 analog, DOTATATE, can be attached to the therapeutic radioisotope ( 177 Lu, typically) and be used for molecular targeted therapy, which is called peptide receptor radionuclide therapy (PRRT) [3]. The uptake of DOTATATE can be observed not only in tumor tissues, but also in normal organs, such as liver and spleen.
This will cause damage to normal organs during PRRT treatment [4]. However, it is still needs further research which targets DOTATATE could binds to besides SSTRs [5]. If we can identify which molecules DOTATATE can combine with and block these molecules during PRRT treatment, it will be possible to reduce the damage of PRRT treatment to normal organs.
Positron emission tomography / computed tomography (PET/CT) permits the measurement of physiological parameters through a dynamic study of tracer dynamics, which is a great advantage over other imaging modalities [6]. In 2019, United Imaging Healthcare released the world's rst total-body PET / CT scanner, named uEXPLORER [7]. With this device, 4D high-de nition dynamic imaging of multi-tissues and organs in the total-body was realized, and it is the only instrument capable of total-body, continuous and dynamic PET scanning at present [8]. The improved sensitivity allows more accurate dynamic imaging of the total-body and dynamic analysis of the physiology of all organs at the same time [9]. Another advantage is that it can obtain high-quality tracer input function from PET images for pharmacokinetic studies, avoiding arterial intubation during blood collection [10]. In addition, the introduction of total-body PET has created an innovative "multisystem biology" framework for the study of human body [11].
The objectives of this study were to map the distribution pattern of 68 Ga-DOTATATE among human organs. Using dynamic total-body PET/CT to calculate standardized uptake values (SUVs) for various organs in the body. Comparing the data with the current knowledge on SSTR2 expression. Moreover, bioinformatics methods were used to compare the correlation between gene sequencing of each organ and DOTATATE uptake, and analyze the proteins that could potentially bind to DOTATATE in addition to SSTR2.

Patient selection
From our cohort of patients, we selected 32 patients (18 men, 14 women, age 18-81 years, median 54 years) who underwent 68 Ga-DOTATATE PET/CT. The patients who were included in this study had no evidence of recurrent disease on conventional imaging performed before the 68 Ga-DOTATATE PET scan.
No tracer-avid lesions were found on the PET scan. All 32 patients met these criteria and informed consent was waived for the study, which was approved by the Institutional Review Board of Renji Hospital and was in accordance with the principles of the 2013 revision of the Declaration of Helsinki.
Imaging Protocol 68 Ga-DOTATATE (2MBq/kg) was administered intravenously, and after a distribution time of 45 minutes, PET/CT images were acquired from the head to mid-thighs with the arms positioned above the head.
Transmission images were acquired to correct for attenuation using a low-mAs CT scan without contrast.

Data Analysis
Two experienced nuclear medicine physicians reviewed the map together, and SUVmax and SUVmean were measured by ROI method. The ROI of different organs was outlined as follows: the ROI of brain, cerebellum, salivary gland, mediastinal blood pool, lung, liver, spleen, prostate and uterus were outlined in the center of the maximum cross section of transverse tomography of each organ, with a diameter of 2 cm. Muscle ROI was delineated in the maximum cross-section of transverse pectoralis major, with a diameter of 1 cm. Skeletal SUV was used to delineate the entire L4 vertebral body at the maximum coronal section. The ROI diameter of kidney, intestinal tract and pancreas (each segment) was 1 cm, and the RENAL ROI was outlined in the renal cortex to avoid urine interference. Intestinal ROI was delineated by small intestine and ascending colon respectively. Pancreas ROI was outlined by head, neck, body and tail segments, and the average was taken as pancreatic SUV. Due to the uneven distribution of radioactivity in intestinal tract and pancreas, the region with the highest concentration of radioactivity should be selected according to visual analysis when sketting ROI, and the interference of surrounding organs such as spleen and kidney to measurement should also be avoided.

Dynamic data acquisition and reconstruction
Considering the high sensitivity of the total-body PET scanner, which allowed for low-dose imaging, the injected dose was a half reduction of the full dose used for routine PET imaging at our department. Lowdose CT was performed for attenuation correction (AC). Then, 60-min dynamic PET scanning was started simultaneously with a bolus injection of 68 Ga-DOTATATE into a vein near the ankle. The images were corrected for radioactive decay, attenuation, scatter, and randoms, and were reconstructed into a 239×239×679 matrix with voxels of 2.85 mm 3 by the list-mode ordered subsets expectation maximization algorithm incorporating TOF and point spread function modeling (OSEMTOF-PSF). For dynamic analysis, the images were divided into 72 frames.

Weighted correlation network analysis (WGCNA)
The co-expression gene and enrichment analysis were applied to WGCNA package of R software, which revealed the correlation between genes [12]. Because the genes with little expression variation usually represent noise, we lter the most variable genes (SD > 1.2) and construct a network. The power of β was set to 5 to ensure a scale-free network. The minimum number of module genes was set to 30. The hierarchical cluster tree summarizes the gene modules of different colors. GO analysis was performed using KOBAS [13].

Gene expression analysis
High-resolution images corresponding to immunohistochemically stained tissue microarrays of different tissue types corresponding to all antibodies analyzed in the present investigation are obtained from the Human Protein Atlas (HPA). The normalized consensus transcript expression levels based on transcriptomics data from the HPA as well as the annotated protein expression levels based on IHC can also be accessed as previously described [14].

PPI network construction
Using STING database to analyze the interaction among the indicated proteins. The tree diagram of PD-L1 related protein network was constructed with the Cytoscape software. The con dence score cut-off was set to 0.8.
Statistical Analysis SPSS 22.0 statistical software and R (https://www.r-project.org/) were used for data analysis. The measurement data conforming to normal distribution was expressed as mean ±SD. Two-sample T test was used for comparison between the two groups, and univariate analysis of variance was performed for comparison between multiple groups. P<0.05 was considered statistically signi cant.

Results
Distribution pattern of 68 Ga-DOTATATE As shown in the maximum intensity projection images, the 68 Ga-DOTATATE was enriched in the bladder, adrenal gland, spleen, kidney, pituitary gland, liver, stomach, intestine, pancreas, salivary gland and thyroid gland (Figure 1). 68 Ga-DOTATATE is excreted through the urinary system, resulting in a large amount of radioactive urine retention in the bladder. In addition, signi cant uptake of 68 Ga-DOTATATE was observed in adrenal gland, spleen, kidney and pituitary gland, while mild uptake was found in liver, stomach, intestinal tract, pancreas, salivary gland and thyroid gland. The average and range of the SUVmax as well as SUVmean for all the organs measured are shown in Table 1. To determine the dynamic process of DOTATATE uptake in each organ, the total-body PET/CT dynamic images were acquired with uEXPLORER ( Figure 2A). And the dynamic uptake curve of multiple organs of the whole body was drawn. The process of dynamic uptake curve is different among organs. In some organs, the uptake of DOTATATE increases gradually, such as pituitary gland, adrenal gland, spleen ( Figure 2A). Some organs, such as the thyroid and pancreas, have an early peak of DOTATATE uptake followed by a gradual decrease in concentration ( Figure 2B). In some other organs, DOTATATE uptake uctuates greatly, and the uptake of DOTATATE is not obvious at the early stage. After a period of rapid uptake, radioactivity reaches the platform. Such organs include stomach, duodenum and small intestine ( Figure 2C).
Somatostatin receptor SSTR2 expression in human organs 68 Ga-DOTATATE is known to be a radiopharmaceutical acting through speci c binding to somatostatin receptor 2 (SSTR2). To explore the reasons for the difference in dynamic uptake of DOTATATE between organs, we analyzed the distribution of SSTR2 expression in human organs with datasets from the Human Protein Atlas [14]. The mRNA expression of SSTR2 in various organs of the human body is showed in gure 3A. SSTR2 expression was found in most tissues of the nervous system, especially cerebral cortex and cerebellum. In endocrine organs, adrenal gland and pituitary gland had a high expression level. In addition, high levels of SSTR2 expression are also found in kidney, Spleen and Stomach. Immunohistochemistry also con rmed that the distribution of SSTR2 protein in organs was basically the same as that of RNA.
In silico screening of potential DOTATATE binding proteins Although DOTATATE is an imaging agent targeting SSTR2, the distribution of DOTATATE in human organs was not completely consistent with the expression of SSTR2. Therefore, there may be other reasons for the concentration of DOTATATE in some organs. It is not clear whether other genes in the human body can bind to DOTATATE. By screening BindingDB, SEA and SwissTargetPrediction databases, and 134 potential binding proteins of DOTATATE were obtained after merging and removing the repeated items ( gure 4A).
To nd the genes related to DOTATATE uptake, we used WGCNA to analyze the genes most closely related to DOTATATE uptake in human organs. After reading and preprocessing data with R software, a total of 19670 expression pro le genes were obtained. In this study, 6542 genes encoding membrane proteins were screened as research objects. Then, with SD>1.2 as the screening criteria, 5755 genes were obtained. When the soft threshold β value is set at 5, the connections between genes are distributed in a scale-free network ( Figure 4B). A total of 36 modules were obtained by WGCNA analysis (Figure 4C and 4D). With GS>0.25 and P.GS <0.05 as the limiting criteria, a total of 800 genes were screened. The targets obtained from the above drug database were intermingled with DOTATATE uptake related genes, and nally eight potential DOTATATE targets were obtained, respectively AVPR1A, EPHA8, EPHB4, OPRL1, SSTR2, SSTR5, ST3GAL1 and NPY1R ( Figure 4E and Table 2).
Next, the protein-protein interaction (PPI) network of these proteins was established with STRING. The PPI network showed that OPRL1 and NPY1R had close relationship with SSTR2 and SSTR5. Whereas AVPR1A, EPHA8, EPHB4 and ST3GAL1 did not interact with SSTR2 or SSTR5 ( Figure 4F).

Discussion
The uptake and distribution of DOTATATE in human organs have been reported in the past [15]. On this basis, this study further explored and more comprehensively described DOTATATE uptake in various organs of the human body. In this study, the uEXPLORER panoramic dynamic PET/CT scanner was used to monitor the dynamic uptake of DOTATATE in all organs of the whole body. By comparing the characteristics of the temporal activity curve of each organ, we found that the dynamic changes of DOTATATE uptake in human organs showed three patterns: one was that the uptake gradually increased and reached a plateau, such as the adrenal gland, pituitary gland and spleen. The uptake of these organs may be due to the presence of DOTATATE target protein expression. The second type of DOTATATE concentration decreases after early rapid uptake, such as in kidney, pancreas and thyroid. The early cause of DOTATATE concentration may be related to blood transport. The third type has a large uctuation range, which uctuates from high to low, such as the stomach, duodenum and small intestine. These organs are all cavity organs, and peristalsis leads to the uctuation of radioactive count.
At present, DOTATATE is known to be an imaging agent targeting somatostatin receptors, among which the binding force with SSTR2 is the strongest [16]. By analyzing the expression distribution of SSTR2 in human organs. We found that the expression of SSTR2 in various organs of the human body was not completely consistent with the uptake degree of DOTATATE. Despite high uptake of DOTATATE in some organs, the expression degree of SSTR2 was not high, such as liver. This suggests that there may be other reasons for DOTATATE uptake in these organs. However, it is not fully discribed in previous studies whether DOTATATE has binding targets other than somatostatin receptors.
Weighted correlation Network Analysis (WGCNA) is a systems biology method used to describe gene association patterns among different samples [17]. It can be used to identify highly covaried gene sets and to identify candidate biomarker genes or therapeutic targets based on the interconnectivity of the gene set and the association between the gene set and phenotype [18]. By using WGCNA, we identi ed the gene set correlated with DOTATATE uptake in human organs, and analyzed the co-expression relationship between these genes. A total of 800 genes correlated with DOTATATE uptake were screened.
In addition, we also predicted and analyzed the proteins that DOTATATE might bind to, by analyzing the molecular structure of DOTATATE. Using BindingDB, SEA and SwissTargetPrediction databases, we obtained 728 potential DOTATATE binding proteins. Among them, 8 genes were found to be positively correlated with DOTATATE uptake by WGCNA, respectively are AVPR1A, EPHA8, EPHB4, OPRL1, SSTR2, SSTR5, ST3GAL1 and NPY1R.
SSTR2 and SSTR5 are known DOTATATE targets. EPH Receptor A8 (EPHA8) encodes a member of the ephrin receptor subfamily of the protein-tyrosine kinase family [19]. EPH and EPH-related receptors have been implicated in mediating developmental events, particularly in the nervous system [20]. Opioid Related Nociceptin Receptor 1 (OPRL1) is a member of the 7 transmembrane-spanning G protein-coupled receptor family, and functions as a receptor for the endogenous, opioid-related neuropeptide, nociceptin/orphanin FQ [21]. Neuropeptide Y Receptor Y1 (NPY1R) belongs to the G-protein-coupled receptor superfamily, and mediates the function of neuropeptide Y, a neurotransmitter, and peptide YY, a gastrointestinal hormone [22]. WGCNA indicates that EPHA8, OPRL1 and SSTR2 all belong to Turquoise module, indicating that there is a co-expression relationship between them. And the PPI network shows that OPRL1 and NPY1R had close relationship with SSTR2 and SSTR5, which may be the reason why they are related to DOTATATE uptake.  Figure 1 Maximum intensity projection (MIP) image showing the distribution of 68Ga-DOTATATE. The important organs are highlighted with numbered arrows: 1, pituitary gland; 2, salivary gland; 3, thyroid; 4, spleen; 5, liver; 6, adrenal; 7, kidney; 8, bladder.