All culturing media ingredients were purchased from Oxoid (Basingstoke, Hampshire, United Kingdom), Macklin (Shanghai, China), and Aladdin (Shanghai, China). Biochemical assay solutions were purchased from Aladdin, Dieckmann (Shenzhen, Guangdong, China), and Shanghai Lingfeng Chemical Reagent co., LTD. (Shanghai, China). I-TASSER server (https://zhanggroup.org//I-TASSER/) was used for the prediction of the 3D structure of the human lactoferrin. Other chemicals used are specified in the methods below.

The rtHLF4, rteHLF1, and rpHLF2 genes were amplified by PCR from the full-length lactoferrin (Accession number: AAA59511) using primers in Supplementary Table S2. These fragment sequences were digested and ligated into pET28b expression vector. All these resulting constructs were confirmed by automated DNA sequencing using universal primers.

The plasmids (pET28b-flHLF, pET28b-rtHLF4, pET28b-rteHLF1, and pET28b-rpHLF2) were transformed into E. coli BL21 (DE3) cultured in lysogeny broth (LB) at 37°C, 220 rpm to OD₆₀₀ = 0.6 followed by 0.5 mM isopropyl β-D-1-thiogalactopyranoside (IPTG) induction at 16°C, 160 rpm for 20 h. Cells were harvested, resuspended in lysis buffer (50 mM Na₂HPO₄, 300 mM NaCl, 10% glycerol, and 1 mM dithiothreitol, pH 7.5), and lysed using EmulsiFlex-C3 (Avestin). Expressed full-length lactoferrin and recombinant lactoferrin fragments (rtHLF4, rteHLF1, and rpHLF2) were purified by immobilized metal ion affinity chromatography nickel-nitrilotriacetic acid (IMAC Ni-NTA). The full-length lactoferrin and lactoferrin fragment rtHLF4 purification was followed by Hiload Superdex 200 size exclusion column chromatography (GE Healthcare). The lactoferrin fragment rteHLF1 and rpHLF2 purification followed by Hiload Superdex 75 size exclusion column chromatography (GE Healthcare). The purified protein was concentrated and stored at −80°C. We use a lipase assay kit (Acmec Biochemical, BC2340) to detect the percentage of LPS in full-length lactoferrin and various lactoferrin fragments. UTI89 was used as the positive control.

Fresh blood was collected from healthy human volunteers (aged between 20 and 40 years old, devoid of any non-steroidal anti-inflammatory drugs for 2 weeks prior) and was mixed with an equal volume of sterilized Alsever solution (2% dextrose, 0.8% sodium citrate, 0.5% citric acid, and 0.42% sodium chloride). The blood was centrifuged for 5 min at 3,000 rpm. The concentrated cell stock was prepared by resuspending cell pellets in isosaline (0.9% w/v NaCl). Cell stock was diluted 10-fold in isosaline to achieve working concentrations. Various concentrations of the flHLF/rtHLF4/rteHLF1/rpHLF2 protein (100, 80, 50, 30, 10, 8, 5, 3, and 1 μM) were prepared in protein buffer, and to each concentration, 1 ml of phosphate buffer, 2 ml of hyposaline, and 0.5 ml of HRBC (human red blood cells) suspension were added. The cell suspensions were incubated at 37°C for 30 min and centrifuged at 3,000 rpm for 20 min. The hemoglobin content in the supernatant solution was estimated spectrophotometrically at 560 nm. Indomethacin (100 μg/ml) was used as the reference standard. The percentage of hemolysis was calculated by assuming the hemolysis produced by the control group as 100%. The HRBC membrane stabilization or protection percentage was calculated using the following formula: percent protection = 100 − [(OD of protein treated sample/OD of control) 100]. All experiments were conducted with the approval of the Institutional Review Board at the Southern University of Science and Technology, under the title of “Evaluation of the safety of therapeutic peptides by hemolysis” with approval number 2020SYG096.

Human non-cancer colon fibroblast CCD-841-CON and CCD-18co, HT29 human colorectal adenocarcinoma cells were obtained from the American Type Culture Collection (ATCC, Manassas, VA, USA). CCD-841-CON and CCD-18co were cultured in Eagle’s Minimum Essential Medium (WISENT, 320–006-CL), and HT29 cells were cultured in McCoy’s 5A medium (CELL COOK, CM 2002). All the cell culture media were supplemented with 1% antibiotics (penicillin/streptomycin) and 10% fetal bovine serum (v/v). All cell cultures were maintained at 37°C in a 5% CO₂ atmosphere.

CCD-841-CON and CCD-18co cell lines were cultured in EMEM medium supplemented with 1% antibiotics (penicillin/streptomycin) and 10% fetal bovine serum (v/v). HT29 cells were cultured in McCoy’s 5A medium supplemented with 1% antibiotics (penicillin/streptomycin) and 10% fetal bovine serum (v/v). These three cell lines were seeded in 96-well plates and incubated for 24 h to allow cell attachment. The cultures were incubated with full-length lactoferrin (flHLF), rtHLF4, rteHLF1, and rpHLF2 till 3-, 24-, and 72-h time points, and untreated wells were also included. The cell viability assay using CellTiter 96® Aqueous One Solution Kit (Promega, USA). Cell culture media was aspirated, and the cells were washed with phosphate-buffered saline (PBS) before incubating the cells with MTS for 1–4 h at 37°C in 5% CO₂. Add 10% SDS to end the reaction. All wells were read at an absorbance wavelength of 490 nm. Relative cell viability was quantified with the fluorescence intensity in the control group considered as 100%. Three independent experiments were performed.

CCD-841-CON and CCD-18co cell lines were cultured in normal EMEM medium with 10% fetal bovine serum (v/v). HT29 cells were cultured in McCoy’s 5A medium, which was supplemented with 10% fetal bovine serum (v/v). These three cell lines were seeded in 96-well plates and incubated for 24 h. Referring to the methods of Venencio’s article, the CCD-841-CON and CCD-18co cells were treated with TNF-α (10 ng/ml) for ROS induction, and all cultures were treated with flHLF, rtHLF4, rteHLF1, and rpHLF2, respectively for 24 h, and TNF-α (10 ng/ml) was used to induce ROS generation and inflammation in CCD-841-CON and CCD-18co (Venancio et al., 2017). The Reactive Oxygen Species Assay Kit (Beyotime, China) was used in ROS assay. Cells were washed twice with PBS before incubating with DCFH-DA for 20 min at 37°C, and the fluorescence intensity of each well was measured using a microplate reader at 488 nm excitation and 525 nm emission. Relative ROS generation was quantified with the fluorescence intensity in the control group considered as 100%. Three independent experiments were performed.

CCD-841-CON, CCD-18co, and HT29 cells were seeded in six-well plates and incubated for 24 h. Cells were treated with lactoferrin flHLF, rtHLF4, rteHLF1, and rpHLF2 for 24 h. In CCD-841-CON and CCD-18co cells, 10 ng/ml TNF-α was used to induce inflammation. Negative (cell only) and positive (TNF-α) controls were also performed. RNA in three wells was extracted using Eastep® Super Total RNA Extraction Kit (Promega, USA) as stated by the manufacturer. For the synthesis of cDNA, 1-μg RNAs of each sample was reversed transcribed to cDNA using the GoScriptTM Reverse Transcription System (Promega, USA), according to the manufacturer’s instructions. Subsequently, 20 ng of cDNAs was mixed with specific primers, and quantitative PCR (qPCR) was performed by using FasStart Universal SYBR Green Master (ROX) (Promega) using the ABI StepOne Plus Real-time PCR System (Applied Biosystems, Foster City, CA, USA). mRNA expression of nuclear factor kappa B (NF-κB), tumor necrosis factor-alpha (TNF-α), interleukin 1 beta (IL-1β), interleukin 6 (IL-6), interleukin 8 (IL-8), and cyclo-oxygenase-2 (COX-2) was analyzed using beta-actin (ACTB) as a reference gene. The primer sequence used in this study is provided in Supplementary Table S1. Data were calculated relative to the control group using the classic 2^−∆∆Ct method.

Protein expression levels were analyzed by Western blot. Cells were seeded in six-well plates and incubated with lactoferrin flHLF, rtHLF4, rteHLF1, and rpHLF2 for 24 h. Cells were washed in PBS and lysed in RIPA buffer (0.6057 g Tris base, 0.877 g NaCl, 10 ml).

Nonident P-40, 5 ml of 10% Na-deoxycholate, 1 ml of 10% sodium dodecyl sulfate, pH set to 7.5, adjusted to 100 ml with H₂O for 30 min before centrifugation and supernatant collection. Subsequently, 20 μl of whole-cell extracts were resolved by sodium dodecyl sulfate-polyacrylamide gel electrophoresis (SDS-PAGE) and transferred to polyvinylidene difluoride (PVDF) membranes. The membranes were blocked with 5% non-fat-milk and then incubated within primary antibodies: ACTB (Abcam, UK, ab213262), NF-ΚB (Abcam, UK, ab76302), TNF-α (Abcam, UK, ab183218), IL-1β (Abcam, UK, ab216995), IL-6 (Abcam, UK, ab233706), IL-8 (Abcam, UK, ab235584), and COX-2 (Abcam, UK, ab179800) at 4°C overnight. The membranes were then incubated with the secondary antibodies (Abcam, UK, ab6721) for 2 h. Blots were visualized by the enhanced chemiluminescence (ECL) method.

All data were presented from three independent experiments. Each experiment has three replicates. Statistical differences between groups were analyzed by Student’s t-test. Statistical significance thresholds are *p ≤ 0.05, **p ≤ 0.01, and ***p ≤ 0.001.

The recombinantly expressed lactoferrin fragments rtHLF4, rteHLF1, and rpHLF2 were based on our earlier published work (Pan et al., 2021). In that study, we observed that the rtHLF4 and flHLF protein could negate the indomethacin-induced hemolytic activity. This preliminary result suggests that on top of the anti-cancer properties, the lactoferrin fragments can further prevent inflammation, further reducing the chances of carcinogenesis of tumor cells.

The soluble recombinantly expressed lactoferrin fragments were successfully purified (Figure 1A) and showed changes in the ferric binding ability supported by the changes in the color of the purified protein (Figure 1B). Based on the iron saturation assay, we found that rtHLF4, rteHLF1, and rpHLF2 showed lower levels of ferric occupancy than flHLF (Figure 1B). rtHLF4, showing a lighter brown hue than flHLF, showed approximately twofold lower in iron saturation, whereas the colorless purified rteHLF1 and rpHLF2 showed less than 5% in iron saturation (Figure 1B inset).

We hypothesize that the iron saturation is due to the truncation of the flHLF, perturbing the ferric binding domains within the lactoferrin structure. In order to understand the changes of rtHLF4, rteHLF1, and rpHLF2 iron saturation profile, we used the I-TASSER modeled structures of these three proteins that showed structural perturbation of the ferric ion-binding pocket resulting from the loss of critical residues within the N- and C-terminus regions of the proteins (Figures 1C–F). Closer investigation of the iron-binding conserved residues (comprising two tyrosines, a histidine and an aspartic acid, and two other carbonate-interacting residues that facilitate metal ion sequestration) showed that all the truncated sequences showed a loss of a single critical residue, resulting in the loss of the iron-binding ability (Lambert et al., 2005). Both rtHLF4 and rteHLF1 retain three conserved residues aside from D63, whereas rpHLF2 retains three conserved residues aside from H249 (Figure 1G, Supplementary Figure S1).

In order to ensure that the purified protein does not influence the inflammatory factors during testing, we found low levels of LPS contamination (<2.5%) from the purified proteins (Supplementary Figure S2).

We employed the indomethacin-hemolytic assay as a preliminary evaluation of the anti-inflammation properties of the four recombinant proteins. The indomethacin-hemolytic assay was previously used to evaluate plant extract’s anti-inflammatory and toxicity properties (Joseph et al., 2013), where indomethacin is used as a measure for membrane cytotoxicity mimicking the conditions of intracellular inflammation (Debnath et al., 2013).

The recombinant flHLF showed minimal inhibitory concentration at 50% (MIC₅₀) of over 80 μM (Figure 2A), similar to the previous reports (Conneely, 2001; Samuelsen et al., 2004; Fedorka et al., 2018). Thus, it is likely that flHLF exhibits anti-inflammatory activities via direct interaction with the host cells and the inhibition of inflammation-causing pathogens (Drago-Serrano et al., 2017). Comparatively, rtHLF4 has an MIC₅₀ of approximately 5 μM, an order of magnitude lower than flHLF (Figure 2B). Although rteHLF1 shows some form of anti-hemolytic properties, the protein could not achieve 50% inhibition despite using up to 100 μM of the protein (Figure 2C). rpHLF2’s MIC₅₀ is above 50μM, showing a slight improvement over the flHLF (Figure 2D).

Our data indicate that all the purified proteins had anti-hemolytic properties, where rtHLF4 had the best activity. Although there is complete inhibition of the indomethacin-induced hemolytic activity, approximately 30% of hemolysis results from erythrocyte instability during the preparation process for the hemolytic assay. Hence, we followed the anti-hemolytic assay with an anti-proliferative assay and a reactive oxygen species (ROS) assay to investigate the protein toxicity and validate the inflammatory properties of these purified proteins.

We tested flHLF and various lactoferrin fragments’ cytotoxicity on non-malignant human colonic fibroblast cells and colorectal cancer cells using a cell viability assay. All the purified proteins do not inhibit the CCD-841-CON and CCD-18co fibroblast cell lines (Table 1, Supplementary Figure S3, S4) but inhibit HT29 colon cancer cells. rtHLF4, in particular, has a higher half-inhibitory concentration at 50% (IC₅₀) in HT29, showing at least 50%–70% higher IC₅₀ values than flHLF (Table 1). This observation suggests that the purified proteins have no cytotoxicity to normal healthy cells but presents anti-cancer properties. The IC₅₀ values of flHLF, rteHLF1, and rpHLF2 tested against HT29 were below 120 μM (Table 1, Supplementary Figures S5A,C,D), whereas rtHLF4 has a lower IC₅₀ value, below 10 μM after 3 and 24 h of treatment (Table 1, Supplementary Figure S5B). These results are consistent with the anti-cancer results from our earlier study (Pan et al., 2021).

The suppression of ROS generation by the purified proteins were used to validate its anti-inflammatory properties. ROS are considered prominent signaling molecules involved in the inflammatory disorder progression, often resulting from oxidative stress/injury. The tumor necrosis factor-α (TNF-α) triggers ROS production via the c-Jun N-terminal kinase activation. In this study, the purified proteins were added to fibroblast or cancer cells treated with TNF-α. rtHLF4 showed the highest inhibition on ROS generation in both the TNF-α-treated fibroblast cells, showing a decrease of over 40% of ROS generated when treated with 10 μM of the protein compared to the untreated fibroblast cells CCD-841-CON and CCD-18co (Table 1). Similarly, the other two proteins (rteHLF1 and rpHLF2) lowered the ROS generation by 10%–20%, where rteHLF1 showed a similar suppression profile to flHLF (Table 1). rtHLF4 also showed the highest suppression of ROS generation in HT29 cells when treated with 0.1 μM of the protein, further supporting the notion that rtHLF4 can prevent colorectal cancer pathogenesis (Table 1). Similarly, flHLF and rpHLF2 could suppress ROS generation to a lesser extent, whereas rteHLF1 showed no inhibition to the ROS generation (Table 1).

Lactoferrin can downregulate ROS levels by triggering the expression of superoxide dismutase (Thome et al., 1997), catalase (Nguyen et al., 2016), and glutathione peroxidase (Wei et al., 2004).

These lowered ROS levels generate a buffer against direct oxidative injuries imposed on the healthy colonic cells (Kruzel et al., 2017). On the other hand, ROS upregulation in cancer cells is attributed to improved anti-cancer activity, resulting in cancer cell apoptosis (Kuete et al., 2015; Zugic et al., 2016). The lactoferrin triggers the buildup of ROS levels within the cancer cell cytosol via the Caspase-8 activation, facilitating the cancer cell apoptosis (Ahmed et al., 2021). This observation concurs with the results of our study, where although the lactoferrin fragments can trigger ROS suppression in cancer cells, the suppression levels are two to three times lower than those found in the healthy fibroblast cells.

In order to fully comprehend the regulatory genes governing the anti-inflammatory properties of purified proteins, we studied the transcriptomic levels of various biomarkers linked to the production of ROS and key inflammation-related genes (Figure 3A). We monitored TNF-α biomarker for its link to ROS production. The inflammation-related biomarkers are divided into the transcription factor NF-κΒ, pro-angiogenesis markers [(Interleukin-6 (IL-6) and Interleukin-8 (IL-8)], and cellular proliferation markers (IL-6 and COX2). We also investigated Interleukin-1β (IL-1β) levels as it regulates NF-κΒ levels. We further validated the cellular protein expression levels of these genes via Western blot (Supplementary Figures S6–S8). Based on the data, we mapped out the role of purified proteins in preventing inflammation (Figure 3B).

Both the colonic fibroblast cell lines treated with the purified proteins showed downregulated levels of TNF-α and the other inflammatory-related genes (Figures 3C–H, Supplementary Figures S6–S8). Of the four different proteins, rtHLF4 showed the best-conferred protection, while the other two proteins showed minimal difference compared to flHLF. Treatment of rtHLF4 on both the fibroblast cells showed approximately 50% lower protein levels of TNF-α (Figure 3C, Supplementary Figures S6A, S7A) and IL-6 (Figure 3F, Supplementary Figures S6D, S7D), 30% lower NF-κΒ (Figure 3D, Supplementary Figures S6B, S7B), and 10% lower IL-8 (Figure 3H, Supplementary Figures S6F, S7F) compared to flHLF treatment. Thus, generally, all lactoferrin variants could suppress ROS production and inhibit inflammation, with rtHLF4 showing the highest efficacy and potency.

Colorectal cancer cell line HT29 treated with flHLF and the various fragments also showed downregulation of ROS production and prevented inflammation via suppression of angiogenesis- and cellular proliferation-related genes (Figure 3B). Compared with flHLF, rtHLF4 treatment showed 20% lower IL-6 and IL-8 expression (Supplementary Figure S8). rtHLF4 also showed 20% lower NF-κΒ expression level (Figure 3D, Supplementary Figure S8B), 30% lower IL-Iβ (Figure 3E, Supplementary Figure S8C), and 60% lower COX-2 (Figure 3G, Supplementary Figure S8E) when compared against flHLF, outperforming the rteHLF1 and rpHLF2. Compared to the healthy fibroblast cells, the lesser suppression of ROS production is essential to facilitate the anti-cancer activity of rtHLF4, whereas the lowering of the IL-1β coupled to the GDF-15 upregulation significantly suppresses NF-κΒ activity (Figure 3B). This results in the inhibition of COX-2 production directly regulated by both NF-κΒ inactivation and the lowered ROS levels.

Earlier findings showed that TNF-α induces ROS production via caspase-8 activation, resulting in the mitochondrial outer membrane permeabilization (MOMP) (Roberge et al., 2014). The MOMP releases ROS and cytochrome C, activating caspase-9 (Kim and Park, 2003). Both caspases-8 and -9 activate caspase-3 resulting in cellular apoptosis (Elmore, 2007). The generated ROS species triggers p50/p65 complex, inducing COX2 expression (Narayanan et al., 2014) (Charalambous et al., 2009). COX-2 overexpression during inflammatory responses often results in the production of prostaglandins involved in colorectal cancer carcinogenesis (Tsatsanis et al., 2006). Our results are consistent with this observation, where lactoferrin suppresses TNF-α expression and prevents cancer carcinogenesis (Kim et al., 2012; Fedorka, 2017).

NF-κΒ activation is the key transcription factor controlling the pro-inflammatory genes, increasing the risk of colitis-associated and other forms of gastrointestinal cancer (Popivanova et al., 2008). NF-κΒ also induces IL-6 through JNK (Lin et al., 2011), playing a role in cancer cell proliferation, stimulating tumor growth and the proliferation of premalignant enterocytes (Becker et al., 2005). Suppression of NF-κΒ via IL-1β (an acute pro-inflammatory cytokine) (Xue et al., 2015) or through GDF-15 ameliorates the inflammation within the host. Our earlier study showed that lactoferrin upregulates GDF15 expression, thereby inhibiting NF-κΒ (Pan et al., 2021). NF-κΒ activation also results in COX-2 upregulation as observed in colonic adenoma and carcinoma (Rehemtulla, 2013) (Charalambous et al., 2009) (Hai et al., 2016). NF-κΒ activation also upregulates IL-8, promoting angiogenesis via JAK2 and STAT3 (Fu et al., 2015; Di et al., 2020). Our results suggest that all the lactoferrin variants, with rtHLF4 in particular, can downregulate NF-κΒ, thus showing anti-inflammation properties and inhibiting tumor growth.