Thyroid nodule is the most common thyroid disease, with an increasing global morbidity in recent years, which is rising with the progress of detection methods (1, 2). Ultrasound is the optimal imaging method for thyroid examination (3). The prevalence of ultrasound has increased the morbidity of thyroid nodules at home and abroad, but may lead to overdiagnosis and unnecessary intervention (4, 5). Since more than 90% of thyroid nodules are asymptomatic benign lesions, only 10% of thyroid nodules are susceptible to thyroid cancer (4, 6). Despite the accuracy and economic applicability, there still exists 2-16% interpreted uncertainty of fine needle aspiration cytology to distinguish benign from malignant thyroid nodules (7). In the past 20 years, the incidental thyroid nodules detected by ultrasound have led to a rapid increase in the morbidity of low-risk papillary thyroid cancer (8). It is estimated that, thyroid cancer will become the four most common malignancies by 2030 (9). Although the morbidity of thyroid cancer has tripled in the past three decades (10, 11), the mortality rate remains relatively stable (12, 13). Papillary thyroid carcinoma (PTC), accounting for 84% of thyroid cancers, is the most common histopathological type of thyroid cancer. Although PTC patients can achieve a favourable prognosis, the 5-year survival rate generally exceeds 97%, and the 10-year survival rate exceeds 90% (14). In addition, there are still a significant number of patients with invasive metastasis, recurrence and iodine-131 resistance, resulting in adverse outcomes. For instance, studies have found that approximately 10-15% of patients may progress to potentially fatal recurrent diseases (15, 16). Accurate preoperative clinical diagnosis has proved difficult due to the lack of specific diagnostic tests for PTC. Therefore, there is an urgent need for a convenient, non-invasive and specific diagnostic method to distinguish PTC from benign thyroid nodules (BTN).

Molecular markers are expected to be used for early detection or screening of thyroid cancer, differentiation of benign and malignant diseases, histological identification, staging and treatment response, diagnosis and prognosis of recurrence (17). The American thyroid Association and the European thyroid Association have published guidelines recommending the detection of molecular markers for uncertain nodules (18). MYC is a human proto-oncogene, including C-MYC, N-MYC and L-MYC, involved in the occurrence, development, invasion and metastasis of tumors. Abnormal expression of MYC in PTC has been reported in a slew of literatures (19). The NDRG family is a downstream regulator of MYC. The AceView database of NCBI showed that, NDRG family members have a NDR-α/β hydrolase folding region and several functional sites, such as phosphorylation sites, acetylation sites, ubiquitin sites, etc. The members of the NDRG family are intracellular proteins, which are composed of 340-394 amino acid residues with amino acid homology of 53-65%. The NDRG family is generally divided into two subfamilies, one composed of NDRG1 and NDRG3 with a homology of 67%, and the other composed of NDRG2 and NDRG4 with a homology of 58% (20). NDRG3 is a downstream regulatory factor of MYC, with a total length of 2588 base pairs. It is mainly expressed in ovary, prostate, testis, brain, spinal cord, thymus, heart and kidney (21). NDRG3 is located on chromosome 20q11.21-11.23 and encodes at least two subtypes, 375 and 363 amino acids, respectively, with an apparent molecular weight of 41 and 40 kDa, respectively (22). However, the relationship between NDRG3 and PTC has not been reported.

Therefore, this study aimed to investigate the differential diagnostic value of NDRG3 in patients with BTN and PTC by detecting the serum and tissues expression level of NDRG3 in normal individuals, BTN and PTC patients, and analyze the relationship between the serum expression level of NDRG3 and clinicopathology of PTC.

There is an ever-growing global morbidity of thyroid nodules with each passing year. Surgery remains the foremost choice of treatment. The high diagnostic rate of uncertain nodules imposes an unbearable economic burden on individuals and society, and significantly reduces the quality of life. Therefore, it is a clinical challenge to perform a surgery on patients with a risk of cancer, and unnecessary surgery should be avoided. It has been pointed out that, for patients with uncertain nodules, the detection of molecular markers can be selected as a supplement. Our current results showed that it can be used as a diagnostic molecular marker for PTC.

We analyzed the preoperative general data and experimental data of serum T3, T4, TSH, TGAb, TPOAb and NDRG3 in 81 PTC patients, 75 BTN patients, 77 healthy volunteers during the same period. The results showed that the median of serum NDRG3 was the lowest in PTC group, followed by BTN group, and the highest in HC group. The median value of serum NDRG3 in PTC group was significantly lower than that in the other two groups. Then we did an immunohistochemical experiment and found that the expression of NDRG3 in the tissues of the three groups was consistent with that in blood, which indicated that the production of NDRG3 was reduced or consumed in thyroid papillary carcinoma and thyroid nodules, so a decrease was detected. We speculate that NDRG3 may be involved in the development of this cellular proliferative disease. It has been found that oxygen and lactic acid regulate NDRG3-mediated lactic acid-dependent signaling pathway. Oxygen negatively regulates NDRG3 expression at the protein level through (proline hydroxylase2) PHD2/(von Hippel Lindau disease) VHL system (23). As members of the NDRG family of oxygen sensors, they exhibit tumor suppressive behavior in various cancers, and their expression is considered to be a good prognostic marker (24). In addition, some studies have found that NDRG3 may be a target for controlling the invasive behavior of hypoxic cancer cells (25). But we do not know its performance in oxygen-rich tissues, we know that cellular hypoxia usually occurs in organs with insufficient blood supply or cells with strong metabolism, while the blood supply of thyroid is abundant, so there is little hypoxia. According to the above point of view, under pathological conditions, NDRG3 may increase in anoxic tissues and decrease in oxygen-enriched tissues. However, studies have found that NDRG3 is highly expressed in gastric and prostate cancer (24, 26), while low expression in breast cancer and oral squamous cell carcinoma (21, 27). From the above studies, it seems that the high expression of NDRG3 in anoxic cancer tissues and the low expression in oxygen-rich cancer tissues seems to be unreasonable.

Recent studies have found that the same member of the NDRG family can have both tumor promotion and inhibition at the same time, which may be determined by the tissue specificity of each member. It has been found that NDRG3 is capable of activating RAF-ERK pathway under hypoxia and accelerating tumor occurrence and development through the accumulation of lactic acid (28). NDRG3 regulates meiotic double-strand DNA break repair by regulating ERK signal pathway in male germ cells (29). In addition, NDRG3 also up-regulate the expression of angiogenic chemokine 43 (CXCL1, CXCL3 and CXCL5), enhance the expression of Angiogen-44, thereby ultimately promoting tumor progression (30). However, in our study, the expression of NDRG3 decreased in two diseases with active proliferation of benign and malignant cells. NDRG3 may be involved in the pathogenesis of PTC as a tumor suppressor gene, the reason of which may be related to the metabolic transformation of cancer cells. It has been found that, NDRG3 is degraded under normoxic conditions, but becomes very stable under hypoxic conditions by binding to lactic acid, even in the case of cell reoxygenation (31). Raquel Guimarães Coelho et al. found the Warburg effect in PTC cell line, indicating enhanced glycolysis and lactic acid fermentation in PTC (32). This may prove that the binding of lactic acid and NDRG3 exists in PTC, so NDRG3 is induced to increase, but combined with more, and eventually detected decrease. In addition, the study found that some BTN may have genetic changes, leading to metabolic changes, similar to thyroid cancer (33), so the NDRG3 detectable in BTN is reduced, but there is no more evidence to support this view. Another study found that NDRG3 increase at the early stage of hypoxia and then decrease during persistent hypoxia (34). Therefore, the same type of cancer may have different manifestations in different stages. Our results showed that the expression of NDRG3 was closely related to extrathyroid extension and lymph node metastasis of PTC. The median value of serum NDRG3 in PTC patients with extrathyroid extension and lymph node metastasis was significantly lower than that in PTC patients without both. Our experiments may suggest that the low expression of NDRG3 is related to the malignant degree of PTC. For example, the description of NDRG3 is inconsistent in the study of breast cancer and liver cancer. The expression of NDRG3 is down-regulated in breast cancer patients, especially in the late stage of the disease (21). In another breast cancer study, the expression of NDRG3 protein is increased, which is related to the invasive biological phenotype and poor prognosis of patients with invasive breast cancer (35). The results of Wang showed that the expression of NDRG3 is up-regulated in hepatocellular carcinoma cells, which reversing the malignant phenotype of hepatocellular carcinoma cells (36). The expression of NDRG3 change the sensitivity of cancer cells to chemotherapeutic drugs, which may be a potential therapeutic strategy for the treatment of liver cancer in the future (37). In conclusion, our study found that the low expression of NDRG3 in PTC is associated with poor prognosis.

Therefore, we speculate that the increased expression of NDRG3 in the general population may indicate the occurrence of hypoxic disease or in the early stage of hypoxia. The decrease or increase of NDRG3 in cancer may be related to the state at that time. If the cell hypoxia persists, lactic acid accumulates, NDRG3 is induced, and the binding with lactic acid increases, the detectable amount decreases. On the contrary, if there is no hypoxia in the cells, NDRG3 will not be pathologically induced and can be expressed normally, and relatively more can be detected in the end. In addition, we speculate that the increase of serum NDRG3 in the general population may play a role as tumor suppressor gene, while a role as carcinogenic gene under the combination of NDRG3 and accumulated lactic acid. These contradictory data may be attributed to differences in tumor micro-environment, tumor type or experimental methods. More specimens are essential to be included in the further study, so as to validate our current findings. There remains several limitations in this study. For example, there is a lack of detailed differentiation of PTC variants and studies of NDRG3 in benign counterparts. We only analyzed a small number of cases, which can be call a limited tissue population study, may be biased, and the experimental design needs further research and improvement.

In Conclusion, the low expression of NDRG3 has a certain differential diagnostic value in patients with BTN and PTC. It is related to the occurrence and development of PTC, and may be a potential marker for diagnosing PTC.

The original contributions presented in the study are included in the article/supplementary material. Further inquiries can be directed to the corresponding author.

The studies involving human participants were reviewed and approved by The Medical Ethics Committee of Gansu Provincial people’s Hospital approved the protocol. The patients/participants provided their written informed consent to participate in this study.

JHW contributed to the conception and design of the experiment and data analysis. JW, CD and LT un contributed to the acquisition of data. JHW drafted the study, which was revised by JL and JQ. All authors contributed to the article and approved the submitted version.