Red blood cell transfusion-induced non-transferrin-bound iron promotes P. aeruginosa biofilms in human sera and mortality in catheterized mice

Summary

Transfusion of storage-damaged red blood cells (RBCs) increases non-transferrin-bound iron (NTBI) levels in humans. This can potentially enhance virulence of microorganisms. In this study, Pseudomonas aeruginosa replication and biofilm production in vitro correlated with NTBI levels of transfused subjects (R²=0.80; p<0.0001). Transfusion of stored RBCs into catheterized mice enhanced P. aeruginosa virulence and mortality in vivo, while pre-administration of apotransferrin reduced NTBI levels improving survival (69% versus 27% mortality; p<0.05). These results suggest that longer RBC storage, by modulating the bioavailability of iron, may increase the risk of P. aeruginosa biofilm-related infections in transfused patients.

Introduction

Human donor red blood cell (RBC) units may be refrigerator stored for up to 42 days by current Food and Drug administration criteria (Dumont and AuBuchon 2008). During storage, RBCs experience progressive damage (i.e., the “RBC storage lesion”), thus impairing their transfusion quality and circulatory lifespan post-transfusion (Zimring 2015). Because standard RBC units contain ~200–250 mg of haemoglobin iron, rapid post-transfusion clearance of significant numbers of storage-damaged RBCs overwhelms the monocyte/macrophage system (Youssef, et al 2018), thereby releasing significant amounts of haemoglobin iron into the circulation in both animal models and humans (Callan, et al 2013, Callan, et al 2021, Hod, et al 2011, Rapido, et al 2017). The rate of iron release can be sufficient to saturate transferrin, the physiologic iron transporter in the circulation, thereby producing non-transferrin-bound iron (NTBI) (Rapido, et al 2017). NTBI promotes proliferation and virulence of specific siderophilic bacterial species; thus, hepcidin-mediated elimination of NTBI may be an important mechanism of host defence (Stefanova, et al 2017). RBC transfusions were shown to exacerbate certain infections in animal models; however, definitive evidence is lacking for RBC transfusion as an independent risk factor for human infection (Prestia, et al 2014).

Pseudomonas aeruginosa, a siderophilic gram-negative bacterium, commonly causes nosocomial infections (de Bentzmann and Plesiat 2011). Biofilm-related infections, an important type of nosocomial infection (Lynch and Robertson 2008), consist of groups of bacteria attached to surfaces embedded in a polymeric matrix; the latter provides protection from antibiotics and the host immune system (Thi, et al 2020). Because iron is a signal for biofilm development by P. aeruginosa (Banin, et al 2005), we sought to determine whether NTBI produced following transfusion of storage-damaged RBCs enhances P. aeruginosa biofilm-related infections.

Materials and Methods

Mice

Outbred, wild-type, jugular vein-catheterized and non-catheterized CD-1 male mice were purchased (Charles River Laboratories, Wilmington, MA) at 8–12 weeks of age. For the blood bank, non-catheterized CD-1 mice were bled aseptically by cardiac puncture into citrate phosphate dextrose adenine (CPDA)-1 solution obtained directly from di-(2-ethylehexyl)phthalate-plasticized, polyvinyl chloride, human primary collection packs (catalogue #4R3611; Baxter, Deerfield, IL) at a final concentration of 14%, as previously described (Hod, et al 2010). Whole blood from 15–40 mice was pooled and leukoreduced using a neonatal high-efficiency leukoreduction filter (Purecell Neo; Haemonetics, Boston, MA), centrifuged at 400xg for 15 minutes, and reduced to a final haematocrit of 60%−80% and haemoglobin of 16–19 g/dl. RBCs were stored in 50-ml Falcon tubes at 4°C for 12–14 days. All procedures were carried out under sterile conditions. On the day of transfusion, 400 μl of fresh (i.e., <24 hours in storage) or stored RBCs (equivalent to 2 human units) were infused retro-orbitally. Concomitantly, mice were infected by tail vein infusion, with P. aeruginosa (1×10⁶ CFU in 200 μl of PBS). Thus, the catheter was not manipulated during the transfusion or infection procedure. A cohort of mice was also injected with 20 mg of human apotransferrin (by intraperitoneal injection before transfusion; Sigma-Aldrich, St. Louis, MO). For NTBI determination in mouse serum, mice were euthanized two hours after transfusion and serum pooled for measurement (N=5 per group). Procedures were approved by the Columbia University Medical Center Institutional Animal Care and Use Committee.

Imaging luciferase activity in vivo

Bioluminescent images were acquired in vivo using IVIS^® Spectrum in vivo imaging system (PerkinElmer, Weltham, MA). Briefly, CD-1 mice were anesthetized with isoflurane and imaged for 1 to 60 seconds at specific time points post-infection. Total bioluminescent flux over regions of interest, expressed in emitted photons per seconds, was quantified using LivingImage computer software (Caliper Life Sciences, Hopkinton, MA).

P. aeruginosa culture

Pseudomonas aeruginosa Xen 41 with a stable copy of the Photorhabdus luminescens lux operon on the bacterial chromosome (PerkinElmer), was inoculated in 30 ml of LB broth and incubated for 16 hours at 37°C in an orbital shaker (aerobic conditions, 250 rpm). The cultures were then washed twice with sterile PBS and pellets suspended in sterile PBS to a concentration of 10⁸ CFU/ml based on spectrophotometric turbidity measurements at 600 nm.

P. aeruginosa co-culture in vitro with human sera or minimal broth

For the experiments in vitro, bacterial suspensions were diluted in serum samples to concentrations of 5 × 10⁶ CFU/ml. Polystyrene, 96-well flat bottom plates (Corning, Inc., catalogue #3591, Corning, NY) were filled with inoculated serum samples (200 μl/well), incubated at 37°C inside humidified plastic chambers to prevent evaporation, and placed in orbital shaker (aerobic conditions, 120 rpm) for defined time intervals. P. aeruginosa was also similarly cultured in minimal broth (2 g of ammonium sulphate, 5 g of glucose, 1 ml of 1 mol/l MgSO4 in 1 L of PBS) spiked to a final concentration of 2, 4, 6, or 8 μM of iron using ferric citrate. To determine planktonic growth, spectrophotometric turbidity at 600 nm (PowerWave XS microplate spectrophotometer; BioTek, Winooski, VT) or luminescence (Synergy H1 hybrid multi-mode reader; BioTek, Winooski, VT) were measured every hour. To determine biofilm formation, contents of each plate were gently removed by washing three times with PBS to remove free-floating planktonic bacteria; wells were then filled with 200 μl of PBS and bioluminescence measured. In addition, plates were incubated for 10 minutes with 0.1% (w/v) crystal violet solution (Fisher Chemical; catalogue #C581–100, Hampton, NH). Excess stain was thoroughly washed away with water and plates were air-dried. Crystal violet was eluted with 30% acetic acid and optical density at 550 nm was determined spectrophotometrically as a measure of biofilm formation. Each measurement was performed in triplicate or sextuplicate.

Human sera

In a prior study (Rapido, et al 2017), 60 healthy adult volunteers were randomized to a single standard, autologous, leukoreduced, packed RBC transfusion after 1, 2, 3, 4, 5, or 6 weeks of storage (n = 10 per group). Blood samples, obtained from a peripherally placed intravenous line, were collected immediately and at 2, 4, 6, 8, 10, 12, 14, and 20 hours after transfusion, as described (Rapido, et al 2017). For this study, samples from 4 subjects per week of storage were randomly selected. Alternatively, human sera were spiked with ferric iron citrate (Sigma-Aldrich, St. Louis, MO) to reach designated transferrin saturations, which were confirmed by measuring serum iron, unbound iron binding capacity (UIBC), and NTBI, as described below. Sera were then used to culture P. aeruginosa strain Xen 41 in 96-well flat bottom plates for 24 hours. Turbidity and luminescence were measured every hour followed by biofilm formation assay, as described above.

Iron-related measurements

Serum iron concentration and Total Iron Binding Capacity (TIBC) were measured using Iron/TIBC Reagent Set (Pointe Scientific, Canton, MI). UIBC was calculated as the difference between TIBC and serum iron. In addition, NTBI was measured by a nitrilotriacetic acid (Sigma-Aldrich, St. Louis, MO) ultrafiltration assay, as described (Rapido, et al 2017).

Statistical analysis

Statistical analysis was performed using Prism 9 (GraphPad software, San Diego, CA) and means ±SD or means ±SEM are shown in the figures, as indicated. In some experiments, area under the curve (AUC) was calculated for each subject over time. Significance was calculated using one-way ANOVA with post-hoc Tukey’s multiple comparison test. Comparison of Kaplan-Meier survival curves was performed using a long-rank (Mantel-Cox) test. Pearson correlations between iron measurements and biofilm bioluminescence was also performed. A p value <0.05 was considered significant.

Results

NTBI enhances in vitro biofilm formation of P. aeruginosa

At first, P. aeruginosa were incubated with human serum spiked with increasing concentrations of iron. Following a 24-hour incubation, P. aeruginosa biofilm production did not increase with increasing transferrin-bound iron; in contrast, it significantly increased when human transferrin was fully saturated and NTBI was present (Fig 1A–B, Fig S1A–B). Further supporting this conclusion, in a similar experiment using minimal broth instead of serum, biofilm production increased at all iron concentrations in the absence of transferrin (Fig S1C–D).

Fig 1. Non-transferrin-bound-iron (NTBI) increases P. aeruginosa proliferation and biofilm production.

Human serum was spiked with increasing amounts of ferric citrate (NTBI and transferrin [Tf] saturation as indicated in panels) and incubated with P. aeruginosa Xen 41 (1 × 10⁶ CFU per well) overnight at 37°C. (A) Bioluminescence in replicate wells (N=6) were measured after indicated incubation times, representing overall bacterial proliferation. (B) Following overnight incubation, wells were washed with saline and bioluminescence of adherent bacteria was measured, which represented sessile bacterial growth. Mean ± SD and statistical significance by ANOVA using post-hoc Tukey’s multiple comparison test of each replicate are shown in the graphs. *p < 0.05; **p < 0.01; ****p < 0.0001.

In our previous study (Rapido, et al 2017), autologous transfusion with human RBCs stored for longer durations increased circulating serum iron, transferrin saturation, and NTBI in some transfusion recipients. This was demonstrated in serum from selected subjects from this prior study (Fig 2A–C) (Rapido, et al 2017). To test the hypothesis that the presence of NTBI after transfusion enhances P. aeruginosa biofilm formation, we incubated sera from these subjects obtained from before to 20 hours after transfusion with RBCs stored for 1–6 weeks with P. aeruginosa. Serum iron, transferrin saturation, and NTBI correlated with biofilm bioluminescence; however, the correlation was strongest with NTBI (Fig 2D–F; R² =0.80; p<0.0001). Furthermore, P. aeruginosa bacterial replication and biofilm production increased after incubation with serum obtained from subjects 4 to 14 hours after transfusion with the most storage-damaged 6-weeks old RBCs (p<0.0001, Fig 2G–H).

Fig 2. In vitro biofilm bioluminescence of P. aeruginosa correlates with NTBI levels in human sera transfused with storage-damaged RBCs.

Serum samples were obtained from a prior study in which healthy human volunteers were transfused with one autologous RBC unit after 1–6 weeks of refrigerator storage (N=4 subjects per week of storage). Serum samples were obtained from immediately after transfusion (time=0) to 20 hours after transfusion. (A) Serum iron, (B) transferrin saturation, and (C) NTBI were measured at defined time points (negative values of NTBI are represented as 0 μmol/l, and transferrin saturation values beyond the linear range are represented as 100%). Linear correlation between biofilm bioluminescence and (D) serum iron, (E) transferrin saturation and (F) NTBI in human sera at 6 hours after transfusion are represented; UND=Undetermined for transferrin values beyond the linear range. Each dot represents a subject. Sera obtained from 0 to 20 hours after transfusion were incubated with bioluminescent P. aeruginosa, and (G) bacterial bioluminescence and (H) biofilm bioluminescence were measured after overnight incubation. Mean ± SEM and statistical significance by ANOVA using post-hoc Tukey’s multiple comparison test of the area under the curve (AUC) for each subject are shown in the graphs. *p < 0.05, ****p < 0.0001. Asterisks without brackets represent significance compared with all other groups

NTBI exacerbates P. aeruginosa central-line associated infection in mice after transfusion with storage-damaged RBCs

To test the hypothesis that NTBI produced following transfusion of storage-damaged RBCs increases the risk of infectious complications in vivo, mice with implanted jugular venous catheters were infected with bioluminescent P. aeruginosa by tail vein infusion (i.e., not through the catheter). The evolution of the infection was recorded for up to 7 days following a concomitant transfusion. A central line-associated infection was confirmed by colonized catheters 24-hours after transfusion and infusion of bacteria (red arrows in Fig 3A). Catheterized mice transfused with storage-damaged RBCs had increased overall mortality (69%) as compared to those transfused with fresh RBCs (13%; Fig 3B).

Fig 3. Transfusion with stored RBCs reduces the survival of catheterized mice infected with P. aeruginosa and improves with pre-administration of apotransferrin.

(A) Representative bioluminescent images of catheterized CD-1 mice infected with P. aeruginosa (1×10⁶ CFU) and transfused with fresh RBCs, stored RBCs, or stored RBCs and pre-infused with 20 mg of apotransferrin (Apo). Mice were imaged daily; catheter location is indicated by the red arrows; (B) Kaplan-Meier survival estimates and (C) bioluminescence over region of interest set around catheter tip over time in cohorts of mice (N=15, 24, and 15 overall for fresh, stored, and stored + apo cohorts, respectively). Statistical significance by log-rank test (B) and ANOVA using post-hoc Tukey’s multiple comparison test of mean ± SEM of area under the curve (AUC) per mouse (C). *p < 0.05.

To investigate whether the increase in mortality after transfusion with storage-damaged RBCs is mediated by NTBI, we administered human apotransferrin to mice before RBC transfusion. Apotransferrin is a potential therapy in this context because it is well tolerated in humans undergoing hematopoietic stem cell transplantation (Sahlstedt, et al 2002), reduces bacterial replication, and enhances the effects of antibiotic treatment in vitro (Ambrose, et al 2019). In the current experimental setting, mice transfused with storage-damaged RBCs have increased NTBI levels post-transfusion, in similar concentrations to those observed in humans (Rapido, et al 2017). By increasing the unsaturated iron binding capacity of serum, apotransferrin administration at the time of RBC transfusion reduced circulating levels of NTBI, but not total iron (Fig S2). In addition, apotransferrin administration improved the survival of mice transfused with storage-damaged RBCs (27% mortality; p<0.05; Fig 3B).

Finally, no time-related differences in the bacterial bioluminescent flux in a region of interest selected over the tip of the catheter were detected among the various mouse cohorts (Fig 3C).

Discussion

Iron is an essential element for eukaryotic and prokaryotic organisms, and bacteria have several mechanisms for acquiring iron necessary for their survival, proliferation and virulence (Cassat and Skaar 2013). Transition to sessile growth and biofilm formation allows bacteria to survive under inhospitable conditions (Lesouhaitier, et al 2019). RBC transfusion represents a life-saving treatment for several haematological conditions and in clinical settings (e.g. intensive care units) where the requirement of blood transfusions is common; however, this procedure increases the availability of iron that can be used by micro-organisms. This has the potential to increase the side effect profile of RBC transfusions.

We previously found that transfusion with storage-damaged RBCs increases the level of circulating serum iron, transferrin saturation, and NTBI (Rapido, et al 2017) (Fig 2A–C). In this study, the increased proliferation in vitro of P. aeruginosa and biofilm formation in some human sera (Fig. 2F–H) suggests that the NTBI produced following RBC transfusion may be associated with increased infectious complications. This observation is further supported by the increased mortality in mice with P. aeruginosa central line-associated infections, transfused with storage-damaged RBCs (Fig 3B).

In prior studies in vitro (Stefanova, et al 2017), siderophilic organisms only grew in plasma when NTBI was present. Furthermore, by controlling NTBI levels, hepcidin selectively protected against siderophilic extracellular pathogens in other mouse models in vivo (Stefanova, et al 2017). Although we did not utilize hepcidin analogues to test their potential to mitigate biofilm-related infections in this model of P. aeruginosa catheter-related bloodstream infection, our results with apotransferrin to eliminate NTBI following transfusion suggests that hepcidin analogues could also be effective in limiting biofilm-related infections.

In our in vivo imaging experiments, no differences in bacterial bioluminescence flux were observed (Fig 3C). This might be because the system detecting the total luminescent flux in vivo is not sensitive enough to distinguish small bioluminescent differences that may have an outsized impact on the outcome of infection. Alternatively, NTBI may affect other aspects of biofilm development, such as maturation or dispersion, and not have a significant impact on the quantity of bacteria adhering to the catheter surface. Indeed, iron regulates expression of multiple bacterial genes, including extracellular virulence factors, such as exotoxin (Visca, et al 2002).

In summary, this study provides evidence that NTBI produced following transfusion of even a single, storage-damaged RBC unit could enhance biofilm development in humans. A limitation of the current study is that only a single strain of P. aeruginosa was utilized to test the hypothesis and does not address whether NTBI is critical for enhancing biofilm-related infections with other pathogens. Nonetheless, P. aeruginosa is responsible for many cases of nosocomial (de Bentzmann and Plesiat 2011) and deadly infections in immunocompromised subjects (Thi, et al 2020). Further investigation is needed to translate these findings to human clinical settings where RBC transfusions are common and may be an important risk factor for biofilm-related infectious complications.