Authors’ reply

We sincerely thank Huynh and Ong for their insightful remarks¹ on our recent publication² comparing serum proteomic profiles of patients with mild (MAP) and severe acute pancreatitis (SAP).

The first and most important concern raised by the authors of the letter referred to the exclusion of patients with moderately severe AP (MSAP) from proteome assessment. Notably, our study was designed as an exploratory, hypothesis‑generating analysis, being the first step of a broader research project. Our aim was to search for proteins that may point to a severe course of AP already in the first days of the disease. Due to methodological, organizational, and financial constraints, this preliminary study could only involve a limited number of samples. Therefore, we decided to include samples from selected patients with clearly distinct AP severity levels (ie, MAP vs SAP), with the hope of detecting a number of significant differences in protein abundance, so that we could identify at least several novel and interesting candidate markers. Currently, our aim is to verify the clinical accuracy of the selected candidate markers for early identification of SAP in larger studies involving more numerous and diverse groups of AP patients. These ongoing validation studies—in line with the comment—also include individuals with MSAP.

Further remarks underlie the possible influence of clinical factors other than AP severity on the differences in protein abundance discovered in our study. We agree that comorbid diseases and treatment factors might have influenced our results (in fact, this was mentioned as one of the limitations in our article). The study, having been restricted to a small number of samples, did not allow for reliable adjustment for confounders. Hopefully, this limitation will be overcome in the larger validation studies. On the other hand, the strength of proteomics lies in the possibility to discover differences in protein abundance that are unexpected and not anticipated by preliminary hypotheses. Many human proteins exhibit multiple functions,³ and proteome analysis may inspire further research leading to a discovery of novel functions of known proteins. Serum proteome analysis, especially after depletion of the most abundant serum proteins, enables the detection of not only functional serum proteins but also the proteins leaking from cells (in the context of our study, these include damaged pancreatic cells, infiltrating cells recruited in inflammation, distant cells affected by systemic inflammation in severe AP, and other, unrelated cells). Our study compared samples obtained from patients with MAP and SAP, collected in the same way and at the same time intervals early after the onset of AP symptoms, and did not include patients with known active neoplasms. Therefore, it seems more possible that the differences in protein abundance were influenced by the extent of local (pancreatic) and / or systemic inflammation than by skin contamination or undiagnosed malignancy, as implicated by Huynh and Ong,¹ although such sources cannot be definitely excluded at this stage. With respect to immunoglobulin domain–containing 4 (VSIG4), we may hypothesize that increased serum levels of this protein in SAP may be due to shedding of VSIG4 from cellular membrane of tissue macrophages following activation of toll‑like receptors (TLRs; eg, TLR4 by damage‑associated molecular patterns),⁴ which may occur within the pancreas or within distant tissues as part of systemic inflammatory response. Unfortunately, we cannot offer a similar hypothesis regarding S100A7 yet, although its expression has been recently documented in human neutrophils.⁵

Finally, Huynh and Ong¹ posed an interesting question about the influence of antibiotic treatment on the differences in protein abundance on the second day of hospital stay vs the day of admission. Although the widespread use of antibiotics in SAP is not supported by current practice guidelines, they are commonly used in Poland due to a fear of septic complications. However, the clinical data summarized in Supplementary material, Table S1 (including data on antibiotic use) were collected over the entire hospital stay.² Even though some patients whose samples were used in our study did not receive antibiotics from the beginning of hospitalization, but later during the course of AP, there are inevitable differences in the treatment of mild and severe cases that might have indeed influenced serum proteome, in addition to the changes induced by the dynamics of pancreatic damage and the inflammatory process itself. This issue is hard to address in clinical studies, and may require well‑controlled animal experiments to provide pathophysiological explanations. However, it may be feasible to include treatment modalities as confounders in the ongoing validation studies.

To conclude, we are grateful for the opportunity to further discuss our research, and we hope that our validation studies will at least partially address the important issues raised in the commentary.