CSF is a severe swine disease in China , resulting in substantial economic loss. However, we are not fully understand the viral pathogenesis, multiplication, distribution, tissue tropism and excretion in pigs. In this study, TaqMan-qPCR was used to detect the viral RNAs in various internal organs, including heart, duodenum, brain, stomach, skeletal muscle, bladder, oesophagus, spinal cord, kidney, ileocaecal valve, rectum, tonsil, liver, lung, jejunum, ileum, submandibular lymph node, spleen, pancreas, inguinal lymph node, mesenteric lymph node and blood at different time points p.i. Using these experimentally infected pigs as a model, we determined the dynamic distribution and the viral tissue tropism. At the same time, an E2 monoclonal antibody WH303-based IHC test was used to verify the quantitative results and the same trend were observed.
The virus load of CSF was almost highest in lymphoid tissue and blood. Interestingly, we found that the ileum and the pancreas showed high viral colonization density. Meanwhile, the heart, duodenum, brain, stomach and skeletal muscle had relatively low viral content. It is surprising that the tonsil, which is anticipated to contain the most abundant viral load, had only moderate viral content.
Many researchers have done similar studies using different strains and different target tissues [8–12]. Previous studies have shown that the CSFV antigen was not detected in heart tissue at any time and that the highest levels of antigen detection were found in the tonsils, spleen and pancreas using IHC test . Considering these findings, we can explain these differences as follows. First, The clinical signs of CSF are extremely variable and it may be confused with many other diseases. The age of the animal, virus virulence and dose lead to distribution differences among acute, chronic and prenatal forms . It may be possible to detect positives for highly virulent strain and high dose groups earlier than those of lower virulence strains or doses because of more rapid transport to the tissues and higher viral load delivered. Second, the detection methods such as virus isolation, AC-ELISA , in-situ hybridization and IHC , RT-PCR, RT-nPCR [16, 17] and real-time PCR [3, 18] have different sensitivities. Real-time PCR is believed to be more sensitive than the other assays. Third, we collected samples daily from day 1 to day 8 post infections, when the pigs were close to agonal stage, which is different from other models, and may be responsible for distribution differences. Other sample collection times may miss the viral multiplication peak.
Here, we determined the viral load in 21 different organs and blood, including the digestive system, respiratory system, circulatory system and nervous system. This provided us with larger samples so that we could obtain better understanding about the distribution of viral infection.
Relative quantitative RT-PCR, as described by Livak and Pfaffl [19, 20] and termed the Livak model and the Pfaffl model, respectively, is widely used to quantify mRNA transcription [21–23]. However, in this study, the dynamic distribution was determined using colonization density as the final determination index. This technique can be applied to assess the relative viral expression to mRNA expression within tissues. Our study is one of the first to use this approach. This concept was recently reported for infectious bursal disease virus . To achieve this goal, an internal control gene (ACTB), which has the highest stability across tissues with consistent expression, was needed  to normalize the initial RNA levels between different samples. Selection of high-quality reference genes is of great importance for the interpretation of data generated by real-time qPCR. In our experiment, the CSFV and ACTB were optimized to have equal amplification efficiency, which is suitable for the Livak model, and it provided reliable results. Recently, some studies have showed that it is better to use multiple internal reference genes for relative quantification . Nevertheless, we only used ACTB as an endogenous gene; if we added another gene such as GAPDH, 18sRNA or HPRT as complementary internal control, this may have provided different results. We consider that, despite the different internal control, the final results will not be greatly affected for the following reasons. First, all the candidate reference gene are expressed at constant levels relative to viral amplification. Second, the copy number of reference genes per cell is much less than the RNA copy numbers of the infected virus.