This study sought to identify a blood-based biomarker of DILD and was the first study to evaluate the usefulness of SFN as a DAD marker. In fact, until now, the relationship between SFN and DAD-type interstitial pneumonia has been unclear, and, to the best of our knowledge, no study has reported the detailed behavior of SFN in the blood. In a validation study using our established SFN ELISA, we observed that the serum levels of SFN were specifically increased in patients with DAD. Thus, our findings on SFN were reproduced in samples from an independent cohort. Moreover, the biomarker performance of SFN for discriminating DAD was superior to those of known biomarkers such as KL-6 and SP-D, which are currently used in clinical settings. Additionally, our data suggests that SFN may be useful not only for the diagnosis of drug-induced DAD but also for the diagnosis of idiopathic DAD. We also showed that the levels of SFN were increased not only in the serum, but also in autopsy specimens and the BALF from patients with DAD. Serum SFN levels were significantly correlated with respiratory parameters. In addition, upregulation and extracellular release of SFN were observed in response to the activation of p53 in pulmonary epithelial-derived cells (A549). These findings suggest that serum SFN levels might reflect the degree of lung injury in patients with DAD; thus, this biomarker would be useful for the diagnosis of DILD patients with DAD as well as for the monitoring of their response to treatment.
SFN (14-3-3σ) belongs to the 14-3-3 family, together with six other isoforms, namely β, γ, ε, ζ, η, and τ. This family is a group of highly evolutionarily conserved proteins with a molecular mass of 25–30 kDa, expressed in all eukaryotes and involved in the modulation of various cellular processes via the binding to phosphorylated proteins22. While most 14-3-3 isoforms are expressed in various normal tissues in a ubiquitous manner, SFN is expressed specifically in stratified epithelia23. In fact, immunohistochemistry staining images of SFN included in the Human Protein Atlas database (https://www.proteinatlas.org) show that SFN is expressed permanently and abundantly in the squamous epithelium of the skin and esophageal tissues; on the other hand, SFN is only very weakly expressed in the normal alveolar epithelium. However, our immunohistochemical analysis using DAD autopsy specimens revealed that SFN was expressed mostly in the basal cells of the bronchioles and some alveolar type II cells that tended to show squamous metaplasia, which is generally observed in the mid to late phase of advanced DAD. Although SFN has been reported to be associated with lung adenocarcinoma24,25, no such association was observed in the serum samples of our lung cancer patients (Fig. 2 and data not shown). Instead, we would like to emphasize that the marked increase of serum SFN levels occurred specifically in the patients with DAD.
The pathological changes of DAD proceed consistently through discrete but overlapping phases: the early exudative phase, the mid proliferative (organizing) phase, and the late fibrotic phase. They reflect the global mechanisms of wound repair, and are thought to be involved in cell cycle regulation and apoptosis26. Denudation and apoptosis of alveolar epithelia may be important early features of acute lung injury. Guinee et al. reported that the expression of the transcription factor p53 as well as the expression of its transcription targets WAF1 and BAX were increased in type II pneumocytes in the lung of a patient with DAD27,28. In addition, Bardales et al. showed that decreases in cell proliferation and increases in apoptosis were frequently observed in the lung tissues of patients with DAD, with both changes tending to be stronger in more severe cases, while apoptosis was undetectable in non-DAD patients29.
It is also worth noting that SFN is a direct transcriptional target of p5330. In the current study, we showed that A549 cells, which were derived from type II alveolar epithelial cells, were able to release SFN in vitro via apoptosis in the context of p53 activation (Fig. 6). These findings suggest that SFN may be upregulated by p53 activation when lung epithelial cells are damaged in patients with acute lung injury such as DAD, and extracellularly released via apoptosis, resulting in increased blood SFN levels. In particular, release of SFN into the alveolar epithelium, which is the main field for gas exchange from the lungs to the blood, may contribute significantly to the increase in circulating SFN levels.
There have been many studies on the intracellular function and biological activity of SFN. The expression of SFN has been shown to be induced in a p53-dependent manner in response to DNA damage, and the progression of the G2/M phase of the cell cycle is arrested30. In addition, SFN is a marker of the differentiation of epidermal keratinocytes. An analysis based on skin biopsy samples revealed that the expression of SFN in keratinocytes in the squamous epithelium was stronger in areas close to the epidermis; moreover, an experiment using cultured cells found that the upregulation of SFN was closely related to a reduction in cell proliferation and the progression of cellular senescence23.
Moreover, in addition to intracellular SFN, the role of extracellular released-SFN has also been reported. Previous studies have demonstrated that the expression of SFN increases when keratinocytes are exposed to ultraviolet rays, and that keratinocyte-released SFN affects dermal fibroblasts via paracrine actions, facilitating the expression of matrix metalloproteinase (MMP)-131–33. In addition, in keratinocytes, the extracellular expression of SFN protein is markedly increased compared to the intracellular expression33, consistent with our data showing that the extracellular SFN levels following p53 activation were much higher than the intracellular levels (Fig. 6b). The above-mentioned studies also suggested that keratinocyte-released SFN may be associated with skin wound healing and suppression of skin fibrosis31–33. Based on our results, we believe that the lung epithelia-released SFN may also be associated with pulmonary damage and remodeling. However, the mechanistic link between the increase in SFN in blood and the pathophysiology of DAD requires further study.
DNA-damaging agents (DDAs) such as cisplatin, carboplatin, and bleomycin may potentially activate p53 in cells. However, in this study we observed no significant difference in the serum SFN levels between DILD patients who were administered DDAs and those who were administered other types of drugs (Supplementary Fig. 3). Importantly, these findings show that the measurement of serum SFN can detect DILD patients with DAD irrespective of the therapeutic agents used, including DDAs. On the other hand, high SFN levels were rarely observed in patients with lung cancer (or other types of cancer) (Table 3 and Supplementary Table 10). Therefore, when using serum SFN for monitoring the onset of DAD in oncology patients undergoing chemotherapy, it may also be necessary to measure the SFN levels prior to drug administration.
Several limitations of this study bear mention. First, although we were able to replicate the current findings, the number of patients enrolled was small. Second, overall, as only the serum levels were examined, the findings of this study do not provide sufficient evidence to indicate a direct relationship between the pathology of DAD and SFN. Indeed, immunohistochemistry data was obtained using autopsy samples from patients with DAD (not from patients assessed for serum SFN levels). Third, high levels of SFN were observed in certain DILD patients with the NSIP or OP pattern, as well as in some lung cancer patients. Therefore, to clarify the reason behind the high levels of SFN in patients without DAD, it is necessary to analyze a larger number of samples. Lastly, this was a retrospective study. Although we demonstrated that SFN increases in patients with DAD, irrespective of either disease types or suspected causative drug types, we cannot completely rule out a potential bias in the selection of patients. Currently, we are collecting serum samples chronologically and prospectively from patients prescribed EGFR-TKI and immune checkpoint inhibitors for future analysis. The analysis of these samples should shed light on the continuous changes in the levels of SFN that accompany the onset of DAD.
In summary, in this study, we identified a protein biomarker for the diagnosis of DAD. SFN showed DAD-specific diagnostic performance, unlike known biomarkers such as KL-6 and SP-D. Since DILD and AE-IIP patients with DAD are likely to have a poor prognosis, it is necessary to make a diagnosis for such patients at the earliest period. Our data suggest that measuring the serum levels of SFN helps to distinguish DAD from other types of DILD, such as OP and NSIP, facilitating the diagnosis of DAD. Therefore, adopting this method may help in the determination of treatment plans for patients with DAD (in addition to the current diagnostic approach based on HRCT findings). Moreover, because SFN is also effective in discriminating the acute phase from the recovery phase, this protein may be useful for monitoring the response to treatment.