IFITM3 Is Elevated in BAL and Plasma Following Infection in ICU Setting Authors: Srabani Das, Carleen R. Spitzer, Paul Consiglio, Matthew C. Exline, Anasuya Sarkar Published: 2026 ABSTRACT ABSTRACTInterferon‐induced transmembrane protein 3 (IFITM3) is an antiviral protein with emerging roles in regulating inflammatory pathways. Although much is known about the role of IFITM3 in mediating host antiviral defense, its contribution to systemic immune responses during critical infectious illnesses remains less understood. Here, we provide evidence that IFITM3 levels were significantly higher in cell free BAL fluid in septic patients with pulmonary infections and in circulating plasma extracellular vesicles (EVs) from critically ill septic patients compared to critically ill non‐septic patients and healthy individuals. This study provides the first evidence of IFITM3's association with sepsis induced inflammatory responses and bacterial pulmonary infections, raising its potential as an agent of dysregulated immunity during systemic infection. Introduction Strong innate immune responses can limit infection in its early stages, while hyperactive inflammatory responses to microbes can drive morbidity and mortality in septic patients. The type I interferon response is a critical inflammatory innate immune mechanism that upregulates hundreds of interferon‐stimulated genes and promotes inflammation. IFITM (interferon‐induced transmembrane) proteins belong to a family of small homologous proteins that have been described as restriction factors for viral infections. Among them, interferon (IFN)‐induced transmembrane 3 protein (IFITM3) is a 15 kDa protein that acts primarily as an antiviral protein against a diverse array of enveloped viruses including Influenza A, Ebola filovirus, and SARSCoV‐2 by inhibiting virus‐cell membrane fusion reactions and restricting cellular infection [1, 2, 3]. Deletion of IFITM3 results in increased vulnerability to viral infections. For example, IFITM3 global −/− mice succumb to sublethal doses of influenza virus [3], and in a large meta‐analysis of over 3500 human subjects, IFITM3 polymorphisms were associated with severe influenza virus infections and hyper‐inflammatory responses [4]. Although much is known about the roles of IFITM3 in mediating host antiviral defense, it has never been studied in the context of critically ill patients with presumed infection. Methods Patient Cohort This is a convenience sample of previously collected specimens. In our first cohort (“BAL Cohort,” Table 1), BAL (bronchoalveolar lavage) samples were obtained from 34 patients (> 18 years) admitted to the general hospital wards or to the medical intensive care unit (MICU). These patients were identified by their treating physicians as having presumed pulmonary infection and bronchoscopy was clinically indicated to aid with further diagnoses and treatment plans. Prior to their clinically indicated bronchoscopy, patients were approached to enroll in our research study as per the approved institutional protocol (IRB 2016H0009). The diagnostic bronchoscopy with BAL was performed by the treating physicians in accordance with institutional standards and 1 mL of excess BAL fluid was used for research purposes. 10 additional anonymous health BAL fluid was used as controls. TABLE 1Patient cohort demographics and diagnoses.DemographicsBAL Cohort Non‐Septic (n = 10)BAL Cohort Septic (n = 24)Critical Illness Cohort Non‐Septic (Plasma) (n = 13)Critical Illness Cohort Septic (Plasma) (n = 26)Age58.5 ± 16.457.3 ± 11.365.8 ± 16.555.5 ± 14.9Gender (M)8/10 (80%)9/24 (38%)6/13 (46%)13/26 (50%)On vasopressors0/10 (0%)8/24 (33%)10/13 (77%)18/26 (69%)Mechanically ventilated1/10 (10%)19/24 (79%)13/13 (100%)26/26 (100%)Acute respiratory distress syndrome (ARDS)1/10 (10%)10/24 (42%)0/13 (0%)13/26 (50%)WBC9.7 ± 5.822.2 ± 38.713.5 ± 6.814.3 ± 8.0Positive BAL culture6/10 (60%)13/24 (54%)3/13 (23%)7/26 (27%)Other positive culture2/10 (20%)13/24 (54%)4/13 (21%)12/26 (46%) Note: Demographic information regarding the two unhealthy patient cohorts. The BAL cohort includes patients with suspected pulmonary infection who were admitted to the hospital. This cohort underwent bronchoscopy at the discretion of their treating pulmonary, critical care physician and excess BAL fluid was sent for testing. Patients were later subdivided into “non‐septic” and “septic” subgroups based on the Sepsis‐3 criteria. The Critical Illness Cohort patients were enrolled within 24 h upon ICU admission. Similar to the BAL Cohort, patients were determined to be septic or not based on Sepsis‐3 criteria. Patient cohort demographics and diagnoses. Note: Demographic information regarding the two unhealthy patient cohorts. The BAL cohort includes patients with suspected pulmonary infection who were admitted to the hospital. This cohort underwent bronchoscopy at the discretion of their treating pulmonary, critical care physician and excess BAL fluid was sent for testing. Patients were later subdivided into “non‐septic” and “septic” subgroups based on the Sepsis‐3 criteria. The Critical Illness Cohort patients were enrolled within 24 h upon ICU admission. Similar to the BAL Cohort, patients were determined to be septic or not based on Sepsis‐3 criteria. In our second patient cohort (“Critical Illness Cohort,” Table 1), 40 critically ill patients were enrolled upon admission to the medical ICU (IRB 2017H0099), resulting in a yield of 39 samples with sufficient material for analysis (the last patient's sample had insufficient plasma and was excluded from further analysis). A separate group of 13 healthy volunteers, (age, year, median [interquartile range] of 36[21–67]; sex male [%] 6[46]) were enrolled as controls. Ten milliliters of blood were collected within 24 h of admission to the MICU as part of a larger study following IRB2017H0099‐approved informed consent. Plasma was subjected to sequential centrifugation followed by ultracentrifugation for extracellular vesicles (EVs) isolation. Monocytes from the peripheral blood mononuclear cells after plasma separation were further isolated via Corning lymphocyte separation medium followed by purification with anti_CD14‐coated magnetic microbeads loaded onto a MACS column following the manufacturer's recommendations (Miltenyi Biotec, Bergisch Gladbach, Germany). Monocytes and plasma EVs were lysed and normalized for protein and analyzed for IFITM3 levels. All patients were adjudicated using Sepsis‐3 criteria [5]. ARDS was identified according to the Berlin definition [6], and survival was defined as survival to hospital discharge. Participant charts were reviewed by two blinded physicians. IFITM3 levels were correlated with clinical diagnoses and severity of illness in both cohorts. Extracellular Vesicles (EV) Isolation EVs were isolated from patient plasma samples via differential centrifugation following our previously validated approach [7, 8] and characterized by CD63 expression [7, 8, 9]. Briefly, plasma were spun at 2000 g for 5 min followed by 16 000 g for 5 min to remove cells and debris. Cell free fractions were further spun at 100 000 g for 1 h at 4°C to yield EV pellets. EVs were resuspended in RIPA lysis buffer and quantified for total protein concentration by DC Lowry assay (Bio‐Rad Laboratories, Hercules, CA). Immunoblot Protein normalized plasma EV and monocyte lysates as described above were subjected to immunoblot and analyzed for IFITM3, IFITM1/2 (Proteintech, Rosemount, IL) and CD63 (Novus Biologicals, CO), followed by appropriate secondary antibody conjugated to HRP. β‐actin (MP Biomedicals, CA) and αβ‐tubulin (Cell Signaling, MA) were used as the loading control for normalization quantification. Enhanced chemiluminescence (GE HealthCare) was used to visualize protein bands and densitometry was performed with ImageJ‐win64 software. Statistical Analysis Data are represented as the mean ± standard error of the mean (SEM) from at least three independent experiments. Statistical comparison between groups was done using Student t‐test where p < 0.05 was considered significant. All patient sample comparisons were performed with ANOVA, followed by post hoc comparison analysis using Tukey–Kramer HSD. p < 0.05 was considered to represent statistical significance. A Wilcoxon test was performed to evaluate differences in BAL fluid between patients. Data were analyzed with JMP statistical software. Results In our pulmonary infection cohort undergoing BAL (“BAL Cohorts,” Table 1), of the 34 individuals diagnosed with infection, 24 (71%) were determined to be septic. A total of 19 patients had positive BAL cultures, 13 in the septic group and 6 in the non‐septic group. BAL from 10 anonymous healthy donors was used as controls. IFITM3 levels in cell‐free BAL samples were significantly higher in septic patients compared to non‐septic patients with diagnosed infection (p < 0.005) (Figure 1A). IFITM1 and IFITM2 were not detected in the infected patient BAL by immunoblot. No IFITMs [1, 2, 3] were detected in BAL fluid from healthy donors. IFITM3 is elevated in BAL and plasma in critically ill patients. BAL fluid (BAL Cohort) (A) and plasma samples (Critical Illness Cohort) (B) from patients and healthy donors were measured for IFITM3 levels. (A) BAL samples were processed via centrifugation at 16000Xg to remove cell debris. Equal volumes of soluble cell‐free BAL were then loaded for immunoblot analysis for IFITM3. IFITM3 levels were normalized to the total BAL volume obtained from each patient for analysis. Due to limited volume of available BAL samples, we were unable to obtain enough EVs after sequential centrifugation for analysis. (B) Plasma EVs from Critical Illness Cohort and healthy donors were protein normalized and analyzed for IFITM3. Based on our observation that IFITM3 levels were significantly higher in BAL fluid from septic patients, plasma samples from critically ill ICU patients with sepsis (“Critical Illness Cohorts,” Table 1) were analyzed for IFITM3 levels. Although there was heterogeneity in IFITM3 levels detected among patient plasma, circulating EV‐encapsulated IFITM3 was significantly elevated only in critically ill septic patients (p < 0.005). Little to no IFITM3 was detected in healthy or non‐septic critically ill patients (Figure 1B). We were also unable to detect other IFITMs, IFITM1 or IFITM2 in these EVs. Monocytes from these patient samples were also analyzed for IFITM3 levels. Although there was a significant increase in released IFITM3 in EVs from critically ill septic patients, no significant difference between cellular IFITM3 levels was observed between healthy, non‐septic or septic patient monocytes. The most commonly encountered infectious organisms in our patient cohorts were Staphylococcal species (34%) followed by E. coli (19%). Only two patients were identified to have active influenza viral infection. There was no statistical difference observed between groups in terms of age, gender, SOFA scores, lactic acid, procalcitonin concentrations or survival. However, our prospective analysis was not sufficiently powered to assess the association of IFITM3 with important clinical outcomes including length of stay, organ dysfunction or mortality. Nonetheless, these data demonstrate that systemic circulation of IFITM3 in EVs occurs during critical infectious illnesses requiring hospitalization. Discussion This study provides the first preliminary evidence of IFITM3's association with bacterial sepsis‐induced inflammatory responses and pulmonary infections, raising its potential as an agent of dysregulated immunity during systemic bacterial infection. IFITM3 and other members of the IFITM family have been described as potent antiviral agents. However, the role of circulating IFITM3 in the context of nonviral sepsis and immune dysregulation is unknown. The significance of elevated levels of IFITM3 in BAL and plasma EVs in critically ill septic patients remains to be determined. The fact that IFITM3 levels were significantly higher during sepsis suggests that IFITM3 levels in part reflect immune dysregulation mediated organ injury. Notably, elevated IFITM3 levels were observed in cell‐free BAL fluid from patients with sepsis, supporting a role for local pulmonary immune activation in addition to the systemic response captured in plasma EVs. There are several limitations of this study. It is important to emphasize that both cohorts represent convenience samples, with inherent heterogeneity in severity of illness, infection type, and clinical trajectory. Moreover, within the BAL Cohort, patients underwent bronchoscopy (and were subsequently enrolled at the study) at a time determined by their treating physician and not on admission to the hospital or ICU as was the case in the Critical Illness Cohort. Comparisons between BAL and plasma EV data are therefore exploratory, as the samples were obtained from distinct patient populations. These findings should be interpreted as hypothesis‐generating observations rather than definitive conclusions. To our knowledge, prior studies have not directly evaluated IFITM3 levels in the context of sepsis or bacterial infection; however, elevated type I interferon signaling and interferon‐stimulated gene expression have been reported in septic patients, providing biological plausibility for the elevations we observed [10]. Also, because we did not assess the kinetics of IFITM3 levels throughout the progression or resolution of disease, we cannot mechanistically link IFITM3 to disease pathogenesis. Furthermore, future work will also need to address the relative contribution from different cells including immune and tissue cells to circulating IFITM3, kinetics of IFITM3 throughout evolution of disease and its impact on clinical outcomes. Author Contributions A.S. conceptualized the project and directed the study and designed experiments in consultation with all authors. C.R.S. and M.C.E. directed the acquisition of patient samples, analysis and interpretation. S.D. conducted experiments, P.C. assisted S.D. in sample processing. A.S. wrote the manuscript with editorial input from all authors. Funding This work was supported by HHS | NIH | National Heart, Lung, and Blood Institute (NHLBI), NIH/NHLBI HL160947. Disclosure The authors have nothing to report. Conflicts of Interest The authors declare no conflicts of interest.