Article Text

Download PDFPDF

Original article
Global microarray profiling identified hsa_circ_0064428 as a potential immune-associated prognosis biomarker for hepatocellular carcinoma
  1. Qiaoyou Weng1,
  2. Minjiang Chen1,
  3. Maoquan Li2,
  4. Yong-Fa Zheng3,
  5. Guoliang Shao4,
  6. Weijun Fan5,
  7. Xi-Ming Xu3,
  8. Jiansong Ji1
  1. 1 Key Laboratory of Imaging Diagnosis and Minimally Invasive Intervention Research, Affiliated Lishui Hospital of Zhejiang University/The Fifth Affiliated Hospital of Wenzhou Medical University/The Central Hospital of Zhejiang Lishui, Lishui, China
  2. 2 Department of Interventional and Vascular Surgery, The Tenth People’s Hospital of Shanghai, Tongji University, Shanghai, China
  3. 3 Cancer Center, Renmin Hospital of Wuhan University, Wuhan, China
  4. 4 Department of Radiology, The Cancer Hospital of Zhejiang Province, Hangzhou, China
  5. 5 Medical Imaging Center, Sun Yat-sen University Cancer Center, Guangzhou, China
  1. Correspondence to Dr Xi-Ming Xu, Cancer Center, Renmin Hospital of Wuhan University, Wuhan, China; doctorxu120{at} and Dr Jiansong Ji, Department of radiology, Affiliated Lishui Hospital of Zhejiang University/the Fifth Affiliated Hospital of Wenzhou Medical University/The Central Hospital of Zhejiang Lishui, Lishui 323000, China; lschrjjs{at}


Background Increasing evidence has shown that circular RNAs (circRNAs) are involved tumourigenesis and metastasis of hepatocellular carcinoma (HCC); however, progression about its function in HCC is relatively slow. Here, we aimed to investigate whether plasma circRNAs could reflect the tumour-infiltrating lymphocytes (TILs) in HCC tumour tissues and serve as prognosis biomarker for HCC.

Methods Tissue samples of patients with HCC were subjected to immunohistochemistry staining against CD8 to examine the TILs. Then, we investigated the expression profile of circRNAs by microarray between plasma of patients with HCC with high TILs and low TILs, and the differentially expressed circRNAs were validated with qRT-PCR. Statistical analysis was performed with SPSS software and GraphPad Prism.

Results We have demonstrated that patients with HCC with high TILs exhibit a significant better overall survival, suggesting clinical outcome could be predicted by TILs. Global circRNA microarray between plasma of patients with HCC with high TILs and low TILs successfully identified six differentially expressed novel circRNAs. Among them, the expression of hsa_circ_0064428 was significantly reduced in patients with HCC with high TILs but increased in patients with low TILs. Moreover, hsa_circ_0064428 was negatively correlated with patient’s survival, tumour size and metastasis.

Conclusion These findings together imply that hsa_circ_0064428 could be considered as a potential HCC prognosis biomarker. Future in-depth research is required to further illustrate the involvement of hsa_circ_0064428 in HCC tumourigenesis and metastasis.

  • hepatocellular carcinoma
  • circular rnas
  • tumor infiltrating lymphocytes
  • hsa_circ_0064428

Statistics from


Tumourigenesis and metastasis are dynamic and complicated biological processes that involve altered microenvironment and immunity.1 2 Tumour-infiltrating lymphocytes (TILs), a type of lymphocytes distributed around tumour cells, are the most crucial indicator of immune response.3 4 CD4+ and CD8+ T lymphocytes are the well-characterised TILs with distinctive function.5 6 Multiple studies have suggested that high TILs in the tumour microenvironment corresponds to improved clinical outcomes such as improved disease-free survival or overall survival (OS) in different cancer types.7–9

Hepatocellular carcinoma (HCC) is one of the most frequently occurring malignancies and leading cancer mortality worldwide.10 Liver cirrhosis and fibrosis, chronic inflammation by hepatitis B virus infection or alcohol intoxication are the well-recognised risk factors.11–15 It is discouraging that majority of patients with HCC are diagnosed at advanced stages with metastasis, and only a limited number of them are qualified for curative treatment.10 16 17 Thus, resolving the early diagnosis of HCC is still urgent and challenging.14 18 19 Antibodies against immune checkpoint molecules promoted the proliferation of CD8+ and CD4+ TIL as well as cytokine production in response to stimulation of tumour-associated antigen in HCC.20 Enrichment of CD4+ and CD8+ TIL predicts longer OS and time to recurrence for HCC after liver resection.21 Although TILs have been suggested as an indispensable factor for evaluating cancer prognosis, its clinical significance in HCC requires more profound investigations and solid evidence.

Circular RNAs (circRNAs) are covalently closed non-coding RNA produced by back-splicing of precursor mRNAs.22 23 At the beginning, circRNA was considered by-products of splicing; however, later they were proven stable, tissue specific and cell type specific, expressed throughout the whole developmental stages.24 25 Recently, next-generation sequencing and novel biological methods have identified numerous novel circRNAs that function as sponges for miRNA or proteins.26–30 Altered circRNA levels lead to aberrant downstream target gene expression that may result in tumourigenesis including epithelial–mesenchymal transition.31–34 Notably, preliminary studies have illustrated regulatory capacity in HCC pathogenesis and metastasis, for instance hsa_circ_0005075, hsa_circ_0004018, hsa_circ_0085154 and so on.35–38 Based on the previous findings, there is no doubt that circRNAs are intimately associated with the pathogenesis of HCC and attractive diagnosis and prognosis biomarkers; however, the precise pathophysiological importance of circRNAs to HCC remains largely elusive.23 39


Patients and clinical samples

A total of 120 participants who were diagnosed with HCC were included in this study. These patients contained an equal number of men and women at different tumour stages. This study obtained approval from hospital ethics committee and written consent from patients. Both plasma and tumour tissue samples were collected from each patient, and the tissues were reviewed by two pathologists independently. Detailed clinical, pathological and molecular characterisations of these patients were collected accordingly.

Microarray analysis

Total RNA from patient plasma was extracted and purified by RNeasy mini kit (QIAGEN, Germany) following instructions. RNA quality was comprehensively assessed by Qubit Fluorometer (Thermo Fisher Scientific, USA). CircRNA microarray profiling and bioinformatics analysis was collaborated with Shanghai Biotechnology Corporation.

qRT-PCR quantification

Total RNA was extracted using TRIzol reagent (Invitrogen, USA) from patient plasma samples following instructions. RNA concentration, integrity and purity were measured by Nanodrop 2000 (Thermo Fisher Scientific). The cDNA was synthesised by the SuperScript Reverse Transcription System (Invitrogen). qPCR assay was performed in triplicate using the KAPA SYBR FAST qPCR kit (Sigma-Aldrich, USA) on an ABI 7900 real-time PCR platform (Thermo Fisher Scientific). The specificity of amplification (prime sequence for circRNAs are shown in online supplementary table 1) was confirmed by melting curve analysis. The relative expression of different circRNAs were calculated by the 2−ΔΔCt method and represented as the means±SD.

Supplementary file 1


Formalin-fixed paraffin-embedded (FFPE) tissues were sectioned at the thickness of 5 µm and placed on coated slides. Sections were first deparaffinised with xylene and dehydrated in a series of ethanol with incremental concentration. Antigen retrieval was carried out by heating at 100°C for 20 min in citrate buffer (pH 6.0) (Biogenex, USA). Then, the slides were subjected to staining with mouse CD8 monoclonal antibody (Dako, Denmark) for 1 hour at room temperature.2 After washing three times with TBS containing 0.05% Tween (TBS-T), the slides were further incubated with HRP-labelled secondary antibody (Dako, USA) for 1 hour at room temperature. Followed by another three times of wash with TBS-T, signals were developed with 200 µL DAB–peroxidase substrate solution. Ultimately, the slides were mounted and examined under a microscope.

Statistical analysis

All statistical analysis was performed with SPSS software V.22 and GraphPad Prism V.6.0 (GraphPad, USA). Data were compared by Student’s t-test as required. P value in bold indicates statistical significance. The survival curves were calculated using Kaplan-Meier method and evaluated by a log-rank test as previously described.40 Statistical significance was defined by p values below 0.05.


High TILs predict better survival in HCC

We first investigated the clinical characteristics of TILs by analysing the expression of cytotoxic T-lymphocyte marker CD8 in FFPE specimens of selected patients with HCC. Interestingly, we observed a high level of TILs in 46.7% of HCC specimens and undetectable TILs in the other 53.3% of specimens (figure 1A). There was a discernible difference in TIL level; therefore, we categorised patients with HCC into TILs high group and TILs low group to explore the association of TILs with clinical response in HCC.

Figure 1

High level of tumour-infiltrating lymphocytes (TILs) predicts better hepatocellular carcinoma (HCC) survival. (A) Immunohistochemistry of CD8+ TILs in patients with HCC. There was a differential level of CD8+ T cells among specimens examined, and a representative picture of high TILs and low TILs is shown (magnification: ×1, ×5, ×10).2 Kaplan-Meier plot of overall survival between high TILs group and low TILs group of patients with HCC. High TILs level was associated with significant improved overall survival of patients with HCC (p=0.034).

Kaplan-Meier survival curves exhibited that patients with HCC with high TILs in tumour tissues exhibited a significant better OS (p<0.05) (figure 1B). The probability of OS of patients with HCC with low TILs at 24 months and 60 months was 0.27 and 0, while the probability of patients with HCC with TILs was 0.81 and 0.29, respectively. This encouraging finding suggests that the percentage of TILs was associated with the OS of patients with HCC after treatment.

Microarray profiling identified significantly deregulated circRNAs

To understand the molecular mechanism and search for HCC prognostic biomarkers, we performed microarray analysis comparing the global circRNA expression profile between high TILs and low TILs patient plasma samples (figure 2A). Microarray had identified 369 differentially expressed circRNAs in total, including 139 upregulated and 230 downregulated in patients with HCC with high TILs (figure 2B,C, fold change >1.5, p<0.05). We chose 10 novel circRNA candidates by conscientiously reviewing more stringent parameters such as normalised FFPM, variation between replicates and referring to established circRNA database and publications. Among these 10 candidates, five were upregulated in patients with high TILs, namely hsa_circ_0065964, hsa_circ_0011386, hsa_circ_0044172, hsa_circ_0010882 and hsa_circ_0052350; meanwhile, the other five downregulated  circRNAs included hsa_circ_0076558, hsa_circ_0064428, hsa_circ_0055538, hsa_circ_0006060 and hsa_circ_0046568 (figure 2D).

Figure 2

Microarray profiling identified 10 potential circular RNAs (circRNAs). (A) Boxplot of microarray data after normalisation.2 The volcano plot depicts the expression difference of circRNAs between patients with hepatocellular carcinoma (HCC) with high tumour-infiltrating lymphocytes (TILs) and low TILs. Coloured points refer to significant deregulation according to fold change and p value, and blue and red represent downregulation and upregulation, respectively. (C) Differentially expressed circRNAs between patients with HCC with high TILs and low TILs. Each column represents a plasma sample, and each row corresponds to a circRNA. ‘Red’ indicates elevated expression level, and ‘green’ indicates decreased expression level. (D) Heatmap of 10 deregulated circRNAs between patients with HCC with high TILs and low TILs. ‘Red’ and ‘blue’ indicates higher and lower expression level, respectively.

Validation of differentially expressed circRNAs by qRT-PCR

We verified the expression of these 10 differentially expressed circRNAs in HCC plasma by qRT-PCR assay. Expectedly, 6 of the 10 selected circRNAs were concordant with the microarray results (figure 3). Compared with patients with HCC with low TILs, the expression of four circRNAs (hsa_circ_0065964, hsa_circ_0011386, hsa_circ_0044172 and hsa_circ_0010882) and two circRNAs (hsa_circ_0064428 and hsa_circ_0055538) was detected significantly upregulated and downregulated in patients with HCC with high TILs, respectively (figure 3A–F, p<0.005). On the contrary, no significant expression difference was found in 4 of the 10 selected circRNAs (hsa_circ_0052350, hsa_circ_0055538, hsa_circ_0006060 and hsa_circ_0046568), which was discrepant with microarray results (online supplementary figure 1).

Supplementary file 2

Figure 3

qRT-PCR validation of the differential expressed circular RNAs identified by microarray between patients with HCC with high tumour-infiltrating lymphocytes (TILs) and low TILs using plasma samples. GAPDH was used for normalisation.

Expression of hsa_circ_0064428 negatively correlated with HCC survival

We therefore questioned whether these differentially expressed circRNAs could predict TILs and act as prognostic markers for patients with HCC. Surprisingly, Kaplan-Meier survival curves revealed that the expression of hsa_circ_0064428 correlated perfectly with HCC survival (figure 4). The OS at 24 months and 60 months was 0.36 and 0.16 in patients with HCC with high hsa_circ_0064428 compared with 0.72 and 0.27 in patients with HCC with low hsa_circ_0064428 (figure 4). There was no clear correlation between survival and the remaining five differentially expressed circRNAs. Consequently, the expression of hsa_circ_0064428 was negatively correlated with the OS of patients with HCC, suggesting that hsa_circ_0064428 could be considered as an independent HCC prognostic biomarker.

Figure 4

Kaplan-Meier survival curves. Strong correlations between hsa_circ_0064428 expression and overall survival of patients with hepatocellular carcinoma (p=0.033) were observed. Log-rank test was used.

To confirm the prognostic role of hsa_circ_0064428, we further examined the clinicopathological parameters in a total of 120 patients with HCC, of which 60 patients had high hsa_circ_0064428 expression and the other 60 had low hsa_circ_0064428 expression. Consistently, 52/60 patients with HCC with low hsa_circ_0064428 expression present high TILs, while 48/60 patients with HCC with high hsa_circ_0064428 expression present low TILs (table 1). Based on these data, the expression of hsa_circ_0064428 in 83.3% (100/120) patients showed favourable negative correlation with the level of TILs (table 1), suggesting hsa_circ_0064428 could be a potential indicator for TILs in patients with HCC. In line with this, increased percentage of tumour size larger than 5 cm in patients with HCC with high hsa_circ_0064428 expression was observed than in patients with low expression (58.3% vs 26.7%, p<0.01). Furthermore, as shown in table 1, patients with HCC with high hsa_circ_0064428 expression were more likely to develop advanced tumour stages III (56.7% vs 20.0%, p<0.01) and led to metastasis (20.0% vs 10.0%).

Table 1

Clinicopathological features of patients with hepatocellular carcinoma (HCC)


HCC is always associated with high mortality rate mainly because patients are normally diagnosed at advanced tumour stages, thus missing the best opportunities for surgical and treatment.10 19 The early diagnosis and prognosis of HCC is still challenging because of its aggressiveness and recurrence.41 In this study, we have demonstrated that patients with HCC with high TILs in tumour tissues exhibit a significant better OS. This result connects tumour TILs to prognosis of HCC, suggesting that TILs could be considered as an effective indicator for HCC clinical outcome. However, measurement of TILs in reality is invasive and inconvenient. Recently, emerging studies have illustrated that circRNAs play important roles in the initiation and progression of many human diseases including tumours and may function as potential biomarkers for diagnosis and prognosis.42–46 Although a number of literatures have described circRNA involvement in HCC pathogenesis and diagnosis,47–49 whether circRNAs play a role in HCC prognosis is largely unknown. Therefore, we investigated whether plasma circRNAs could reflect the TILs in HCC tumour tissues and serve as a prognosis biomarker.

Overall, we identified six differentially expressed novel circRNAs in patients with HCC with high TILs by combination of global circRNA microarray screening and qRT-PCR validations. Generally, the expression of hsa_circ_0064428 was confirmed significantly downregulated in patients with HCC with high TILs and negatively correlated with patient’s survival, suggesting hsa_circ_0064428 could act as a potential prognosis biomarker. Moreover, we found that hsa_circ_0064428 expression was also associated with tumour size and metastasis, indicating that it was also crucial for promoting tumour growth and progression.

Currently, it is widely recognised that circRNA regulates biological processes by acting as miRNA sponges, thereby manipulating target gene expression at the transcriptional or post-transcriptional level.27 35 50 51 Hence, hsa_circ_0064428 may interact with multiple miRNAs to perturb various tumour-related signalling pathways. Despite future investigations that are required to unveil the precise molecular mechanism, it is incontrovertible that hsa_circ_0064428 shall be proposed as biomarkers for HCC prognosis.

So far, hsa_circ_0064428 has never been studied by any literature and we are the first to identify its relationship with HCC prognosis. CircRNAs have unexampled advantages as diagnostic and prognostic biomarkers considering its universality, specificity and stability.39 52 CircRNAs are widely detected in blood, saliva and body fluid, and their abundance is higher compared with linear RNAs.53 54 Besides, it was reported that circRNA was extremely stable and frequent freeze–thaw does not affect its biological property.55 Recently, studies have indicated that the superiority of circRNAs as prognosis biomarker could reflect the stage characteristics of tumourigenesis.39 48 Therefore, hsa_circ_0064428 is an ideal biomarker candidate for HCC prognosis.

Bioinformatics analysis revealed that hsa_circ_0064428 potentially interacts with multiple miRNAs including hsa-miR-7977, hsa-miR-4530, hsa-miR-4692, hsa-miR-4514 and hsa-miR-4645. A recent study showed that altered expression of hsa-miR-4692 expression could reflect progression of Barrett’s oesophagus to oesophageal adenocarcinoma.56 Meanwhile, hsa-miR-4530 demonstrated involvement in chronic lymphocytic leukaemia.57 In summary, our study demonstrated that the expression of plasma hsa_circ_0064428 is negatively associated with tumour TILs level and survival in patients with HCC, which make it a promising non-invasive prognosis biomarker. Compared with traditional HCC prognosis biomarkers, such as alpha-fetoprotein, hsa_circ_0064428 has theoretically improved sensitivity and specificity. In vivo validations and clinical applications of hsa_circ_0064428 in HCC prognosis are attractive to explore, which might build the foundations for improved cancer treatment and patient management.


  1. 1.
  2. 2.
  3. 3.
  4. 4.
  5. 5.
  6. 6.
  7. 7.
  8. 8.
  9. 9.
  10. 10.
  11. 11.
  12. 12.
  13. 13.
  14. 14.
  15. 15.
  16. 16.
  17. 17.
  18. 18.
  19. 19.
  20. 20.
  21. 21.
  22. 22.
  23. 23.
  24. 24.
  25. 25.
  26. 26.
  27. 27.
  28. 28.
  29. 29.
  30. 30.
  31. 31.
  32. 32.
  33. 33.
  34. 34.
  35. 35.
  36. 36.
  37. 37.
  38. 38.
  39. 39.
  40. 40.
  41. 41.
  42. 42.
  43. 43.
  44. 44.
  45. 45.
  46. 46.
  47. 47.
  48. 48.
  49. 49.
  50. 50.
  51. 51.
  52. 52.
  53. 53.
  54. 54.
  55. 55.
  56. 56.
  57. 57.


  • QW and MC contributed equally.

  • Contributors Conception and design: JJ and X-MX. Provision of study materials or patients: QW, MC and JJ. Collection and assembly of data: ML, Y-FZ and GS. Data analysis and interpretation: WF and QW. Manuscript writing: QW, X-MX and JJ. Final approval of manuscript: all authors.

  • Funding This study was supported by the National Natural Science Foundation of China (no. 81573657), Key Research and Development Project of Zhejiang Province (no. 2018C03024), Zhejiang Province Medical and Health Care Key Project (no. 2016146810), National Natural Science Foundation of Zhejiang Province (no. LQ17H180001), Chinese Medicine Science and Technology Projects of Zhejiang Province (nos. 2015ZA228 and 2016ZA209), Key Discipline Project of Lishui City (2017ZDXK07), National Natural Science Foundation of Zhejiang Province (nos. LQ16H160019 and LQ17H180001), High-Level Talent Project of Lishui City (2014RC01), Science and Technology Department of Lishui City (2017ZDXK07), and Experimental Animal Science and Technology Projects of Zhejiang Province (2016C37101).

  • Competing interests None declared.

  • Patient consent Obtained.

  • Ethics approval Affiliated Lishui Hospital of Zhejiang University.

  • Provenance and peer review Not commissioned; externally peer reviewed.

Request Permissions

If you wish to reuse any or all of this article please use the link below which will take you to the Copyright Clearance Center’s RightsLink service. You will be able to get a quick price and instant permission to reuse the content in many different ways.