Intravenous phage display identifies peptide sequences that target the burn-injured intestine

Abstract

The injured intestine is responsible for significant morbidity and mortality after severe trauma and burn; however, targeting the intestine with therapeutics aimed at decreasing injury has proven difficult. We hypothesized that we could use intravenous phage display technology to identify peptide sequences that target the injured intestinal mucosa in a murine model, and then confirm the cross-reactivity of this peptide sequence with ex vivo human gut. Four hours following 30% TBSA burn we performed an in vivo, intravenous systemic administration of phage library containing 10(12) phage in balb/c mice to biopan for gut-targeting peptides. In vivo assessment of the candidate peptide sequences identified after 4 rounds of internalization was performed by injecting 1×10(12) copies of each selected phage clone into sham or burned animals. Internalization into the gut was assessed using quantitative polymerase chain reaction. We then incubated this gut-targeting peptide sequence with human intestine and visualized fluorescence using confocal microscopy. We identified 3 gut-targeting peptide sequences which caused collapse of the phage library (4-1: SGHQLLLNKMP, 4-5: ILANDLTAPGPR, 4-11: SFKPSGLPAQSL). Sequence 4-5 was internalized into the intestinal mucosa of burned animals 9.3-fold higher than sham animals injected with the same sequence (2.9×10(5)vs. 3.1×10(4) particles per mg tissue). Sequences 4-1 and 4-11 were both internalized into the gut, but did not demonstrate specificity for the injured mucosa. Phage sequence 4-11 demonstrated cross-reactivity with human intestine. In the future, this gut-targeting peptide sequence could serve as a platform for the delivery of biotherapeutics.

Fig. 1

Candidate gut barrier targeting peptide sequences identified during in vivo, intravenous phage screening. DNA sequencing of the recovered phage clones was performed on a small number of clones in order to identify the peptide sequences that were internalized into the intestinal epithelium following burn. Highlighted sequences are found in multiple phage clones recovered from the gut epithelium. (For interpretation of the references to color in this figure legend, the reader is referred to the web version of the article.)

Fig. 2

Candidate peptide sequences were screened in vivo following severe burn to identify sequences which demonstrate burn specificity. An injection containing 1 × 10¹² copies of each selected clone were injected via tail vein into sham or animals 4 h following burn. A total of 3 different peptide sequences (4–1, 4–5, and 4–11) were tested based on results from in vivo biopanning. (A) The amount of phage bearing each candidate peptide sequence that was recovered from the intestine was quantified. Data is expressed as phage particles per mg of intestinal tissue. (B) The relative quantity of phage displaying each candidate peptide sequence that was recovered from burn-injured animals compared to sham was measured.

Fig. 3

Therapeutic window for targeting the burn-injured intestine. Increasing concentrations of phage expressing sequence 4–11 injected intravenously into sham (gray line) and burned animals (black line). For sequence 4–11, the greatest therapeutic window for recovery in the gut between sham and burned animals is seen when 1 × 10¹² phage are injected.

Fig. 4

Phage sequences can be designed to internalize into burn injured intestine tissue. Quantitative PCR of intestinal tissue harvested from animals 4 h after burn injury following intravenous injection of phage. Candidate peptide sequence 4–5 (ILANDLTAPGPR) targets the burn-injured gut to a greater degree compared to sham. *p < 0.05, t-test.

Fig. 5

Peptide sequences identified using phage display in our murine model also bind to human gut tissue. Human small intestine was incubated with phage sequence 4–11 (SFKPSGLPAQSL) then stained with fluorescent antibodies (Alexa 488, green) for visualization using confocal microscopy. Sections of human gut incubated with the secondary and fluorescent antibodies alone (Negative Control) and with the phage containing no peptide sequence (Empty Phage) were used as controls. The binding of our candidate peptide sequence suggests that there is cross-reactivity between human tissue and peptide sequence identified through phage screening in a murine model. (For interpretation of the references to color in this figure legend, the reader is referred to the web version of the article.)