We first analyzed human colon biopsies obtained from cancer patients who underwent treatment with radiotherapy (n=12), untreated colorectal cancer patients (n=11), and those diagnosed with colon ischemia (n=11). Colon ischemia was diagnosed in postoperative histopathology in all cases in the last group. The colonic tissues were stained for HO-1 and γH2AX. We found high levels of HO-1+ Mø in the irradiated colons and ischemic regions of the colon but not in matched non-ischemic tissues from the same patient ( Figures 1A–C ). In ischemic regions, majority of CD68+ Mø expressed HO-1 unlike in matched non-ischemic colons ( Supplementary Figure 1 ). Interestingly, the levels of HO-1 in the epithelium were lower in the irradiated colon and slightly increased in the ischemic regions of epithelium compared to matched control tissue ( Figures 1D, E ). Further, nine patients in the irradiated and ischemic groups exhibited rectal bleeding, defined clinically as presenting with either macroscopic hematochezia or fecal occult blood positivity. This cohort showed higher infiltration of HO-1+ Mø but no change in the epithelial HO-1 levels compared to those without bleeding ( Figures 1F, G ). Since heme is scavenged by Hx, we stained the same tissue with an antibody against Hx ( Figure 1H ). We showed a significant increase of Hx staining in colons from cancer patients who underwent treatment with radiotherapy ( Figures 1H, I ) as well as in the areas of the ischemic colon compared to matched non-ischemic tissues from the same patient ( Figures 1H, J ). No difference in the Hx staining was seen in the cohorts with or without bleeding ( Figure 1K ).

As expected, radiation significantly increased DNA damage (γH2AX) in the colons ( Figures 1L, M ). A non-significant trend towards higher γH2AX staining was observed in ischemic colons compared to matched non-ischemic tissue ( Figures 1L, N ). There was no difference in γH2AX staining between the colons from patients with or without rectal bleeding ( Figure 1O ). These data suggest the importance of HO-1+ Mø and Hx upon the genotoxic or ischemic stress and a possible association of HO-1 expression with bleeding (indicating a presence of free heme) in the colon.

We have also investigated colon tissues from rectal cancer patients (n=8) treated with CRT. CRT regimen was determined individually by institutional tumor board recommendation and consisted of radiation with up to 50.4Gy in conjunction with either 5-FU or capecitabine. We stained the sections of the matched injured (after CRT) and uninjured (before CRT) tissues from the same patients with antibodies against HO-1 and γH2AX ( Figure 2 ). We found a trend of an increased number of HO-1+ Mø infiltrating the colon upon CRT compared to the matched control samples ( Figures 2A, B ) in both patients either with or without rectal bleeding ( Figure 2C ). Importantly, the highest infiltration of HO-1+ Mø was detected in the colons of patients treated with CRT and who also presented with rectal bleeding ( Figure 2C ). A slight increase was seen in patients treated with chemoradiation and without rectal bleeding ( Figure 2C ). This suggests that the effect of CRT on HO-1+ Mø infiltration might be dependent of rectal bleeding. There was no difference in epithelial HO-1 expression between the groups (data not shown). Further, levels of γH2AX varied in the colons of treated patients ( Figures 2D, E ) and did not correlate with the incidence of rectal bleeding ( Figure 2F ). These data support the role of HO-1+ Mø infiltration both in response to radiation as well as chemoradiation ( Figures 1 , 2 ).

Since the number of HO-1+ Mø was strongly increased in the colons of patients with rectal bleeding after anti-cancer therapy, we employed two models of colonic injury in mice lacking HO-1 in myeloid cells (LysM-Cre : Hmox1^flfl ; Cre) and control mice (Hmox1^flfl ). In the first model, we used PHZ to induce systemic hemolysis and colonic damage ( Figures 3A–C ). In the second model, we used a single dose of doxorubicin to promote DNA damage, cell death, and release of heme in the colon ( Figures 3D–I ). We found elevated DNA damage (γH2AX, Figures 3A, C ) and proliferation of epithelial cells (P-Histone H3, Figures 3A, B ) in mice lacking Hmox1 in myeloid cells (LysM-Cre : Hmox1^flfl ; Cre) compared to control Hmox1^flfl mice treated with PHZ ( Figures 3A–C ). These data indicate the importance of clearance of heme by HO-1 in myeloid cells, which may otherwise cause abnormal DNA damage and proliferation responses in the colon. To further investigate the role of HO-1 in myeloid cells, we treated LysM-Cre : Hmox1^flfl and Hmox1^flfl mice with 8 mg/kg doxorubicin (1x, i.v.) (n=4-6 mice/group, including females and males) and harvested colons at days 1 and 5 after treatment ( Figures 3D–I ). No differences were seen between body weights or white blood cells (WBC) levels between the genotypes or treatments ( Supplementary Figures 2A, B ). Interestingly, the lack of myeloid-derived HO-1 resulted in increased platelet (PLT), hemoglobin (Hb), and hematocrit (HCT) levels in untreated mice ( Supplementary Figures 2C–E ). Doxorubicin treatment led to a significant increase in free heme in colons with a peak at 1 day and returned to baseline at day 5 in Hmox1^flfl mice ( Figure 3D ). The levels did not return to basline in LysM-Cre : Hmox1^flfl mice at day 5 after treatment with doxorubicin ( Figure 3D ). The number of infiltrating HO-1+ cells was increased in the colon of Hmox1^flfl control mice at 5 days after treatment with doxorubicin ( Figure 3E ). We found significant decrease in proliferation in the colons of LysM-Cre : Hmox1^flfl in response to doxorubicin at day 1, which returned to a baseline level at day 5 ( Figures 3F, G ). However, no difference in proliferation was seen between the genotypes ( Figure 3F, G ). Doxorubicin treatment led to an increase in DNA damage at day 1 as measured by γH2AX staining in both groups ( Figures 3H, I ), but no difference was seen between the genotypes. These data suggest that the presence of HO-1+ infiltrating cells and free heme in the colon in the GIS model might be important for colonic homeostasis and healing in response to hemolysis/bleeding but not in response to doxorubicin-mediated injury.

By analyzing Geo profiles, we found high levels of Hx mRNA in the mouse ileum and colon compared to other parts of the intestine ( Figure 4A ). Previous studies indicated that Hx ^-/- mice had more severe hemolysis in response to hemolysis induced by PHZ injection than wild type mice (12, 25). To define the role of Hx in GIS models and the connection between increased levels of free heme and colonic damage, we used Hx ^-/- mice treated with PHZ (100 mg/kg, i.p.) ( Figure 4 ). Control and PHZ-treated Hx ^-/- mice had significantly smaller spleens compared to wild type mice ( Supplementary Figure 3A ). The levels of HO-1 and hemoglobin (Hb) were highly elevated in Hx^-/- mice ( Supplementary Figure 3B ) due to elevated hemolysis and free heme levels in response to PHZ treatment ( Figures 4B, C ). Further, free heme promoted DNA damage in the spleen ( Supplementary Figures 3C, D ), similar to our previous report in cancer model (11). We observed a slight decrease in cell proliferation in the colons of Hx^-/- mice treated with PHZ compared to wild-type (WT) mice ( Figures 4D, F ). The decrease in P-Histone H3 (P-HH3) staining corresponded to an increase in γH2AX staining in colons of Hx^-/- mice treated with PHZ ( Figures 4D–G ). We found a significantly higher number of γH2AX+ cells in the colon of Hx^-/- mice basally and after treatment with PHZ ( Figures 4E, G ). We also investigated the efficiency of recombinant Hx (rHx) to revert colonic injury following PHZ treatment in Hx^-/- mice ( Figures 4H, I ). One dose of rHx (2.5 mg/kg, i.v.) did not significantly suppress the DNA damage as assessed by the level of γH2AX staining in Hx^-/- mice treated with PHZ compared to untreated mice ( Figures 4H, I ). These data support a significant role of Hx-mediated protection in the colon in response to hemolysis and elevated free heme levels.

To investigate the colonic injury in the absence of Hx in response to chemotherapy, we used doxorubicin treatment (8 mg/kg, 1x, i.v.) in Hx^-/- mice ( Figure 5 ). We observed increased levels of HO-1+ cells infiltrating into the colon of wild type (WT) mice treated with 8 mg/kg doxorubicin (1x, i.v.) (n=3-4 mice/group) and harvested tissues at 48 hours after treatment ( Figure 5A ). The numbers of HO-1+ cells infiltrating colon ( Figures 5A, B ), as well as Hmox1 mRNA levels in response to doxorubicin treatment, were similar to those seen basally in Hx^-/- mice ( Figure 5C ). Doxorubicin treatment did not increase HO-1 levels in Hx^-/- mice beyond the elevated levels in the untreated Hx^-/- colons. c-MYC mRNA levels remained unchanged in response to doxorubicin in WT mice ( Figure 5D ). However, there was a substantial increase in the expression of c-MYC in Hx^-/- mice, basally and in response to doxorubicin treatment ( Figure 5D ). We showed elevated IL-10 mRNA levels in wild type mice treated with doxorubicin and significantly increased IL-10 expression in Hx^-/- mice ( Figure 5E ). In this model, we also found significantly higher number of γH2AX+ cells in the colon of doxorubicin treated WT mice ( Figures 5F, G ). Further, an increased number of γH2AX+ cells in Hx^-/- mice with or without doxorubicin treatment was observed ( Figures 5F, G ). Doxorubicin did not lead to increased DNA damage in Hx^-/- model beyond what was seen in the WT mice ( Figures 5F, G ). These data were further supported by slightly decreased levels of P-HH3, a proliferation marker, in the colons of WT mice treated with doxorubicin ( Figures 5H, I ). Lack of Hx resulted in lower proliferation in the colonic crypts, but there was no significant difference between WT and Hx^-/- mice treated with doxorubicin ( Figures 5H, I ). The data described above demonstrate the presence of HO-1+ cells infiltrating the colon in the GIS model, whose number/level is induced by doxorubicin. The presence of these cells also suggests that colonic homeostasis in Hx^-/- mice is altered after treatment with doxorubicin.

We have previously shown that radiation associated with bone marrow transplant in Hmox1 deficient mice is associated with persistent DNA damage in the colon (9). To assess the role of heme:Hx in the model of abdominal irradiation (IR), we used local 12 Gy IR (shielded IR) to induce DNA damage in the colon by shielding the upper body of WT and Hx^-/- mice ( Figure 6A ). We showed a slight increase in hemolysis in WT mice treated with abdominal IR ( Figure 6B ), which was significantly elevated in Hx^-/- mice ( Figure 6B ). This increase in free heme in the plasma of irradiated Hx^-/- mice resulted in significantly higher DNA damage and slightly lower proliferation compared to WT mice treated with the same regimen ( Figures 6C–F ). This data suggests a critical role of hemolysis-regulated DNA damage in the colon in response to abdominal radiation.

Previous reports indicated that heme induces hyperproliferation of epithelial cells in the colon (26). Using human colonic epithelial cells (HCoEpiC), we observed an increase in cell proliferation and survival in response to heme ( Figure 7A ). The survival of heme-treated cells was unchanged upon co-treatment with 1 μM doxorubicin, which decreased the total number of viable HCoEpiC cells as expected ( Figure 7A ). In contrast to colonic cells, heme treatment of Mø (RAW267.4 cells) promoted doxorubicin-induced cell death ( Figures 7B, C ). Further, knockdown of Hmox1 in the RAW 264.7 macrophage cell line using stable shRNA against Hmox1 (shHmox1) resulted in decreased survival in response to doxorubicin in the presence or absence of heme ( Figure 7C ). Heme treatment alone at 1-50 μM doses did not induce cell death as we previously reported (27). We also performed real-time PCR in HCoEpiC cells treated with heme at 24 h to validate changes in G4-regulated genes including c-MYC, CCNF, HDAC6, ULK1, FosB as well as a pro-inflammatory cytokine, TNF ( Figure 7D ). Hmox1 mRNA levels after treatment with heme were increased ( Figure 7D ). We found that heme:G4 complexes-regulated genes such as c-MYC and HDAC6 (12) as well as TNF were increased upon stimulation with heme. In contrast, no differences were found in FosB or CCNF gene expression levels ( Figure 7D ). These data indicate a strong impact of free heme on normal colonic cell gene expression profile and their proliferation and survival basally and in response to doxorubicin treatment.

All studies were approved by the IRB committee at the BIDMC. The archival pathology specimens from the Department of Pathology at BIDMC were identified by a search of Pathology records for the period 2018-2021. H&E stained sections of the samples were reviewed by one of the authors (J.G.) and the pathologic diagnoses were confirmed. The medical records were examined to determine the following patient characteristics: age and stage at diagnosis, adjuvant treatment (chemotherapy and/or radiation), and disease status at the time of last-known follow-up. Colorectal biopsies or resections were collected from patients after radiotherapy for primary malignancy or untreated (including rectal, prostate, colon, lung, and esophageal cancer patients) (n=11 control, n=12 irradiated) and from patients diagnosed with intestinal ischemia (n=11) following the IRB protocol. The control group included patients without gastrointestinal symptoms and no history of chemoradiotherapy or intestinal ischemia, who had histologically normal mucosa in biopsies. All samples were fixed in neutral buffered formalin and embedded in paraffin.

Pathology specimens were obtained as paraffin blocks and sectioned at 5 μm. Tissue samples were formalin-fixed followed by paraffin embedding, and immunostaining of 5 μm sections was performed as previously described. The following antibodies were used for human material and mouse tissue staining: γH2AX (Cell Signaling, MA), P-Histone H3 (Cell Signaling, MA), and HO-1 (Enzo Life Sciences; Abcam). Tissues were de-paraffinized and processed for antigen retrieval using high-pressure cooking in citrate buffer. Sections were then blocked for 30 min in 7% horse serum (Vector Laboratories, Burlingame, CA, USA). Primary antibody was then applied to the sections overnight at 4°C. The following day, sections were incubated with biotin-labeled (Vector Laboratories, Burlingame, CA, USA) or fluorescently-labeled (Alexa Fluor488 or Alexa Fluor594; Thermo Fisher) secondary antibody for 1 h at room temperature, followed by VECTASTAIN Elite ABC kit and detection with ImmPACT DAB (Vector Laboratories, Burlingame, CA, USA). All images were captured using a Nikon Eclipse E600 microscope (Nikon Instruments, Melville, NY, USA) or Zeiss Fluorescence Microscope. Protein staining was evaluated by both stain intensity and proportion of positive-staining cells. The intensity scale is as follows: 0 (no pigmentation), 1 (light yellow), 2 (buff), and 3 (brown). The percentage of positive cells was assessed by high power field: 0 (<5% chromatic cells), 1 (5–25% chromatic cells), 2 (26–50% chromatic cells), 3 (51–75% chromatic cells), and 4 (>75% chromatic cells).

All experimental procedures were performed in accordance with relevant guidelines and regulations and were approved by the Institutional Animal Committee (IACUC) at BIDMC. Male and female C57BL/6 wild-type (WT), as well as LysM-Cre : Hmox1^flfl and Hmox1^flfl mice were maintained in our colony as previously described (28). Hx^-/- mice were a kind gift from Dr. Tolosano (29). For PHZ treatment, the mice were treated with 100-150 mg/kg, intraperitoneal injection (i.p.), at 8-12 weeks of age. In a separate group of mice, gamma irradiation was used at a dose of 12 Gy in the supine position. The upper body was shielded using lead shielding plates (Precision X-ray) covering the head, torso, upper extremities, to the low xiphoid. Regions below the xiphoid were irradiated. All the mice were observed daily for any signs of radiation sickness, morbidity, and mortality. Their body weight was recorded daily. Recombinant Hx was obtained from Athens Research and was used at 2.5 mg/kg, i.v. in mice. Animals in the doxorubicin treatment group were treated with a single intravenous (i.v.) dose of 8 mg/kg doxorubicin. Tissues and blood were harvested from the control and treated mice after 12, 24, and 48 h for further analysis. Complete blood count (CBC) was measured using HemaVet 950 analyzer (Drew Scientific, Inc. CT).

Total RNA was isolated from mouse colons or epithelial cells using RNeasy Plus Mini Kits (QIAGEN, Valencia, CA, USA), and cDNA was synthesized using HiFiScript cDNA Synthesis Kit (CWBIO) based on manufacturer instructions. To quantify the gene expression levels, synthesized cDNA and PowerUpTM SYBR Green Master Mix (Applied Biosystems) were used to amplify the target genes. Amplifications were performed on 1µg cDNA using the following primers Table 1 :

The following program was applied: 95°C for 10 min, 95°C for 15 s, 58°C for 55 s, 72°C for 55 s, 95°C for 15 s, 60°C for 1 min, and 95°C for 15 s (steps #2 to #4 repeated for 40 cycles). StepOne software version 2.3 (Applied Biosystems, MA, USA) was used to calculate relative changes in mRNA levels that were normalized to the β-actin levels.

The mouse macrophage cell line, RAW 264.7, was purchased from ATCC and maintained in RPMI media (Life Technologies) supplemented with 10% fetal bovine serum (FBS). HCoEpiC (ScienCell Labs) were purchased and maintained in CoEpiCM full medium as specified by the manufacturer’s protocol (ScienCell Labs). Cells were incubated at 37°C with 5% CO₂.

HCoEpiC cells at passages 3-5 and RAW cells were seeded on 6-well plates at 200 000 cells/well density. Cells were incubated at 37°C with 5% CO₂ overnight to let them adhere. Cells were treated with 5-50 μM heme with or without doxorubicin for 24 hours. Hemin (referred to as heme, Sigma-Aldrich, St. Louis, MO, USA) was prepared by dissolving the powder in 0.1N NaOH and then titrated with 0.1N HCl to biological pH 7.4, followed by adjustment to 10 mM concentration with 0.9% saline. Heme stock was then aliquoted and frozen at −80°C until use; each aliquot was thawed only once. Heme-utilizing experiments were carried out in the dark at various concentrations of 1–50 µM. Doxorubicin hydrochloride (Sigma-Aldrich, St. Louis, MO, USA) stock (2 mg/ml) was stored in the dark at 4°C until use for cell culture treatment 0.5-10 μM.

HCoEpiC or RAW264.7 cells were seeded on a 96-well plate at a density of 20 000 cells/well. Cells were treated with the following reagents: 5-50 μM heme, 0.5-10 μM doxorubicin, and combined treatment with 5-50 μM heme and/or 1 μM doxorubicin for 24 hours. For the control group (six replicates) only, CoEpiCM complete medium was added. After 24 hours, cells were rinsed with 1xPBS (GIBCO, Life Technologies) and stained with crystal violet (Sigma Aldrich) for 20 minutes on a shaker. After staining, plates were washed in water to remove excess staining. Crystal violet-stained cells were dissolved in 10% acetic acid and absorbance was measured at 560 nm using an ELISA plate reader.

Genomic profiles of the mouse intestine were obtained from Geo Profiles (24).

Plasma samples were obtained by centrifuging blood (on EDTA) at 1600 x g for 10 min at 4°C. The plasma was then additionally spun at 16,000 x g for 10 min. The levels of heme/hemoglobin were measured as absorbance at 420 nm. The levels of free heme were measured by Heme Colorimetric Kit (BioVision) as previously described (30).

All data are presented as mean ± standard deviation unless otherwise indicated. Statistical analysis was performed using Student’s t-test or one-way analysis of variance (ANOVA) followed by the post hoc Tukey test using Prism 9.0 (GraphPad Software, San Diego, CA, USA). Differences between groups were rated significant at values of p < 0.05.