Lack of Cellular Inflammation in a Non-human Primate Model of Radiation Nephropathy

Abstract

Inflammation is commonly cited as a mechanism of delayed effects of acute radiation exposure (DEARE). Confirmation of its presence could provide significant insight to targeted use of treatments or mitigators of DEARE. We sought to quantify the presence of cellular inflammation in kidneys of non-human primates that developed acute and chronic kidney injury after a partial body irradiation exposure. We show herein that cellular inflammation is not found as a component of either acute or chronic kidney injury. Other mechanistic pathways of injury must be sought.

INTRODUCTION

The renal injury caused by high dose ionizing radiation has commonly been known as radiation nephritis, the “-itis” suffix indicating inflammation. If correct, this label would have substantial significance because it could imply that mechanisms of injury would be inflammatory, either cellular or humoral, and that mitigators and treatments would need to act via pro-inflammatory pathways. Other normal tissue radiation injuries do have inflammatory components, such as the cellular inflammation that is an acknowledged feature of gastrointestinal radiation injury (MacVittie et al. 2012). The inflammatory components of radiation pneumonitis are known (Johnston et al. 2004), and they account for the generally acknowledged benefit of the use of glucocorticosteroids in treatment of that condition (Hanania et al. 2019).

In a rat model of renal radiation injury, it was noted in histological studies that a cellular inflammatory component was not prominent (Cohen et al. 1996). A literature review showed 10 publications that specifically reported on the lack of inflammation in both human and animal studies (Zuelzer et al. 1950; Fairweather et al. 1960; Cogan and Arieff 1978; Raulston et al. 1978; van Kleef et al. 2000; Mostofi et al. 1964; Maier and Casarett 1964; Phemister et al. 1973; Jaenke et al. 1980).

Nonetheless, inflammation has been proposed as a major mechanism for the delayed effects of acute radiation exposure (DEARE) (Zhao and Robbins 2009). This has major implications for the use of mitigators of radiation injury. The timing of inflammation in an exposed organ would then indicate the targeted timing of mitigators acting to counter-act inflammation. In a previous study of non-human primates (NHP) that underwent 10, 11, and 12 Gy partial body irradiation, inflammatory leukocytes were not found to be a major feature by semi-quantitative evaluation (Parker et al. 2019). We sought to quantify cellular inflammation in our well-characterized non-human primate model of partial body radiation exposure.

MATERIALS AND METHODS

Animals

Non-human primate (NHP) rhesus macaques underwent 10 Gy partial body irradiation (PBI) as previously reported (MacVittie et al. 2012; Farese et al. 2019). Forty-three had 2.5% bone marrow sparing, and 31 had 5% bone marrow sparing. Of those having 2.5% bone marrow sparing, 31 also were treated with two granulocyte colony stimulating factors (GCSF), Neupogen or Neulasta (Farese et al. 2019). Of those having 5% bone marrow sparing, 16 underwent planned rather than for-cause euthanasia before the end of the study. All surviving NHP were euthanized at the end of the planned study period, which was 180 d after irradiation. An additional six were non-irradiated controls. All irradiated NHP were observed for up to 180 d post-irradiation or until they were euthanized.

All NHP were in good health at the start of study and were free of simian immunodeficiency virus, simian T cell leukemia virus type 1, malaria, herpes B virus, and tuberculosis. All animal procedures were performed according to an approved Institutional Animal Care and Use Committee (IACUC) protocol.

Housing, care, food, and water

Animal housing and care were performed in accordance with the Animal Welfare Act and the Guide for the Care and Use of Laboratory Animals. Irradiated animals were single-housed in stainless steel cages at our animal facility accredited by the Association for Assessment and Accreditation of Laboratory Animal Care.

Anesthesia.

Ketamine (Ketaset, Fort Dodge, IA) (10 ± 5 mg kg⁻¹), either alone or in combination with xylazine (AnaSed, Fort Dodge, IA) (1 ± 0.5 mg kg⁻¹), was given intramuscularly (IM) to sedate animals before procedures. Yohimbine (Yobine, Shenandoah, IA) [(0.2 ± 0.1 mg kg⁻¹, IM or intravenously (IV)] was given to reverse xylazine sedation, if required.

Leukocyte growth factor Neulasta® or Neupogen® administration

Neulasta® or Neupogen® (Amgen, Inc., 1 Amgen Center Dr, Thousand Oaks, CA 91320) were dosed (10.2 ± 0.3 µg kg⁻¹) using the previous day’s body weight and injected subcutaneously using four treatment cohorts: Neulasta® was injected once on d 1 (24 h) and d 8 or d 3 (72 h) and d10; Neupogen® was injected once a day starting at d 1 (24 h) or d 3 (72 h) post-irradiation. Dosing of Neupogen® was stopped when an absolute neutrophil count (ANC) ≥ 1,000 cells µL⁻¹ was observed for three consecutive days. Control animals were dosed with an identical volume of 5% dextrose (D5W) (Baxter Healthcare Corporation, Deerfield, IL) in water starting at 24 h post-irradiation.

Radiation exposure and dosimetry

PBI was 10 Gy delivered bilaterally AP/PA with 6 MV, 2 MV average energy linear accelerator-derived photons. Calibration and dosimetry were done as previously reported (MacVittie et al. 2012; Farese et al. 2019). PBI was delivered to midline tissue at an approximate dose rate of 0.80 Gy min⁻¹ with approximately 5% and 2.5% bone marrow sparing.

Individualized medical management

Cage-side observations, clinical observations, analgesics, anti-emetics, anti-ulcerative, anti-diarrheals, antibiotics, anti-inflammatories, anti-pyretics, diuretics, nutritional support, blood transfusions, and enteral or parenteral fluid support were provided as previously reported (MacVittie et al. 2012). Dexamethasone (Butler Schein, Dublin, OH) was administered to animals noted to be in respiratory distress (>80 breaths per min) during daily cage-side observations. NHPs were treated with intramuscular (IM) injections using a planned taper: 1 mg kg⁻¹ twice daily (BID) on the first day of treatment, 0.5 mg kg⁻¹ BID for 3 d, 0.5 mg kg⁻¹ once daily (QD) for 3 d, and 0.5 mg kg⁻¹ every other day (QOD) for three doses (Garofalo et al. 2014).

Euthanasia

Animals were euthanized based on criteria as reported (MacVittie et al. 2012), and all surviving animals were euthanized at 180 ± 10 d post-irradiation. This was in compliance with the American Veterinary Medical Association Guidelines for the Euthanasia of Animals and our IACUC.

Blood chemistry

Blood was taken by saphenous vein venipuncture under ketamine anesthesia just before euthanasia. Blood urea nitrogen (BUN) was determined using commercial kits on an Alfa Wassermann Ace Clinical Chemistry Analyzer (Alfa Wassermann, West Caldwell, NJ). Normal values were confirmed using blood samples from age-matched non-irradiated NHP. Azotemia is an elevation of plasma or serum urea nitrogen (BUN), which is a standard indicator of renal function. Its elevation indicates a loss of renal function; its doubling over baseline values of the naïve NHP indicates a greater than 50% loss of renal function in rhesus macaques (Moghadasali et al. 2014).

Renal histology

Organs procured at necropsy were processed for histology. All samples were immersed in 10% neutral-buffered formalin for at least 48 h, embedded in paraffin, cut in 5-µm sections and stained with Masson Trichrome. Slides were scanned using a Hamamatsu Nanozoomer HT slide scanner (Hamamatsu Corporation, Bridgewater, NJ) and viewed using the Aperio Image Scope software (Leica Biosystems, version 12.3.0.5056, Buffalo Grove, IL). Scoring was done in a masked fashion, without knowledge of the use of Neupogen®.

Acute kidney injury (AKI).

AKI was scored for all specimens. The acute injury score was for tubular injury (none, scattered, or diffuse) glomerular thromboses (none, scattered, or diffuse), and medullary congestion (none, scattered, or diffuse), each grade being 0, 1, or 2 for a maximum score of 6. The six additional NHP that were not irradiated also had scoring for AKI.

Chronic kidney injury (CKI).

CKI was scored for all specimens. This histological scoring was for cysts (none, micro-, or macroscopic), interstitial fibrosis (none, scattered, or diffuse), casts (none, scattered, or diffuse), glomerular thrombosis (none, few, or most glomeruli of 20 examined), glomerulosclerosis (none, 1–2, 3–4, >4 of 20 examined), each grade being 0, 1, or 2, and also for mesangiolysis (none, variable, most, or all glomeruli of 20 examined), each grade being 0, 1, 2, or 3, for a maximum score of 14. Fibrosis was a prominent feature and was thus scored for glomeruli and interstitium. The six additional NHP that were not irradiated also had scoring for CKI.

Cellular inflammation.

We tested the role of cellular inflammation in 55 specimens, which are a subset of the cohort that were stained in a single batch (Table 1). The trichrome stained specimens were used to identify mononuclear leukocytes in the renal parenchyma. Such mononuclear cells are monocytes or lymphocytes. Using the Aperio scan program, two squares, each 300 µm wide and tall, were drawn in the mid-cortex of each kidney. Within each square, a grid of smaller squares was created with 10-µm sides for each smaller square. Mononuclear leukocytes were counted, and a small square was counted positive if one or more mononuclear leukocytes were in it. These are cells with rounded dark nuclei that are likely to be lymphocytes or monocytes. Cells with elongated angular nuclei were not counted as positive. This was done twice, in a masked fashion. Accurate identification of mononuclear inflammatory leukocytes was verified several times using photomicrographs exchanged between two of the authors.

Table 1.

Subgroups of non-human primates used for the present studies. The radiation dose is either none (non-irradiated) or 10 Gy. Three subgroups had 2.5% bone marrow sparing, and two subgroups had 5% bone marrow sparing. Those that were also treated with Neulasta or Neupogen are shown. The numbers of NHP for each subgroup are shown, and the number in parentheses is the number that had specimens available for analysis of mononuclear leukocyte numbers.

Group	Non-irradiated	10 Gy/BM2.5	10 Gy /BM2.5 Neupogen	10 Gy/BM2.5 Neulasta	10 Gy/BM5 scheduled	10 Gy/BM5 for cause
number	6(6)	12(12)	16(12)	16(10)	16(15)	15(0)

Experimental endpoints

The primary endpoints were renal function, as azotemia, the acute and the chronic histological injury score, and the presence of mononuclear leukocytes per unit area. Data were stored and sorted using Excel, version 2010. Statistical analysis was done using GraphPad Prism 8 (San Diego, CA) and included linear regression for the relationship of AKI and CKI scores over time and the relationship of mononuclear leukocytes over time. Analysis of variance was used to compare the numbers of mononuclear cells per each group.

RESULTS

In this cohort of NHP that underwent 10 Gy PBI/BM2.5 or 10 Gy PBI/BM5, there was an early phase of acute and a later phase of chronic kidney injury, as has been reported (Cohen et al. 2017) (Fig. 1). The renal function data for the NHP that underwent 10 Gy PBI/BM5 were previously reported (Cohen et al. 2019). The renal function data for the NHP that underwent 10 Gy PBI/BM2.5 were not previously reported. These may show a greater occurrence of AKI than occurred in the NHP that underwent 10 Gy PBI/BM5. However, conclusive statistical analysis is not possible because of the small numbers and multiple comparison issues.

Fig. 1.

Renal function after 10 Gy PBI, and according to treatment group. The treatment groups are 10 Gy PBI/BM5, no LGF (underwent 10 Gy PBI with 5% bone marrow sparing, had no leukocyte growth factor), 10 Gy/BM2.5, no LGF (underwent 10 Gy PBI with 2.5% bone marrow sparing, had no leukocyte growth factor), 10 Gy/BM2.5, neul 1 (underwent 10 Gy PBI with 2.5% bone marrow sparing, had Neulasta started ond1 after PBI), 10 Gy/BM2.5, neul 3 (underwent 10 Gy PBI with 2.5% bone marrow sparing, had Neulasta started on d 3 after PBI), 10 Gy/BM2.5, neup 1 (underwent 10 Gy PBI with 2.5% bone marrow sparing, had Neupogen started on d 1 after PBI), 10 Gy/BM2.5 neup 3 (underwent 10 Gy PBI with 2.5% bone marrow sparing, had Neupogen started on d 3 after PBI). The normal median BUN is shown as the dotted grey line. There is an acute kidney injury phase within the first 50 d after irradiation, shown as an early rise in BUN, and a chronic kidney injury phase, starting at about 100 d after irradiation, shown as the later rise in BUN. Almost all surviving NHP have chronic kidney injury.

Acute kidney injury (AKI) was found right away after PBI and persisted thereafter (Fig. 2). There was no statistically significant change in AKI score with time after irradiation. In previous studies, we showed that histological AKI appeared to wane with time after PBI. The present studies scored AKI in all specimens. The persistence of AKI-type histological changes beyond 50 d after PBI is not fully explained.

Fig. 2.

Acute kidney injury histological score as a function of time after 10 Gy PBI, and according to treatment group. The grey dotted line shows the acute kidney injury score for non-irradiated NHP kidneys. There is acute kidney injury within a week after PBI, and it persists thereafter. There are no obvious differences depending on individual treatment group. There is no trend of AKI scores with time.

Chronic kidney injury (CKI) was also found soon after PBI. There was a steady rise of the CKI score after that, and all NHP surviving to the end of the study had CKI scores above normal (Fig. 3). There was a statistically significant rise in the CKI score with time after irradiation (r = 0.6, p < 0.0001). This is consistent with a scarring process that starts soon after PBI and is progressive thereafter.

Fig. 3.

Chronic kidney injury histological score as a function of time after 10 Gy PBI, and according to treatment group. The grey dotted line shows the chronic kidney injury score for non-irradiated kidneys. There is chronic kidney injury within 10 d after PBI, and it increases steadily thereafter. The regression line is shown as a grey solid line (r=0.6, p < 0.0001). At study end, 180 d after PBI, all NHP had chronic kidney injury score above normal. There are no obvious differences depending on individual treatment group.

No differences have been found for the renal histology for acute or chronic kidney injury when comparing irradiated NHP that also were treated with leukocyte growth factors (LGF), Neulasta or Neupogen, and those that were not (Cohen et al. 2019). This lack of influence of leukocyte growth factors on the renal injury is apparent in the present cohort.

Mononuclear leukocytes were not a prominent feature. A sample photomicrograph is shown in Fig. 4. The median counts for all groups were all less than 10 per 900 squares, i.e., less than 1% of the parenchyma, and similar to the cell counts for non-irradiated NHP (Fig. 5). There was no difference in the number of mononuclear leukocytes according to group (p = 0.1, by ANOVA). The evolution of mononuclear cell inflammation from time of irradiation showed a possible trend for a modest increase of cell counts over time from irradiation (r = 0.3, p = 0.07), but almost all counts were less than 10 per 900 squares (Fig. 6).

Fig. 4.

Photomicrograph of a trichrome-stained kidney section at approximately 400X magnification. This was obtained from an NHP that had undergone 10 Gy PBI/BM5 45 d earlier. Its BUN was 15 mg dl⁻¹ at time of euthanasia, its histological CKI score was 2, and its histological AKI score was 1. The large green square includes 900 smaller squares, each one having sides of 10 µm. Inspection of two large green squares yielded an average of four small squares containing a mononuclear leukocyte, which is an area of cellular inflammation of 4/900, i.e., less than 1%. One mononuclear leukocyte is indicated by the arrow, a proximal tubule lumen is indicated by a star, and a distal tubule lumen is indicated by an x.

Fig. 5.

Number of squares with a mononuclear cell, arranged according to treatment group. The total number of squares in a single field was 900, and two cortical fields were scored per specimen. There is no difference of the percent area positivity for mononuclear leukocytes depending on treatment group (analysis of variance, p = 0.1).

Fig. 6.

Number of squares with a mononuclear cell as a function of time after PBI. The treatment groups are as in Fig. 1, and their separate symbols are shown. There is an insignificant upward trend of mononuclear cell number with time after PBI (r = 0.3, p = 0.07).

In the present studies and also on analysis of hematoxylineosin stained specimens from this cohort of NHP, neutrophils and eosinophils were scant to absent in irradiated kidneys. Toluidine blue staining also has not shown a significant presence of mast cells (Parker et al. 2019).

DISCUSSION

The present studies extend recent and past data showing little or no role for cellular inflammation as a major feature of radiation nephropathy in non-human primates (Cohen et al. 2017; Parker et al. 2019; Raulston et al. 1978; van Kleef et al. 2000). This includes consideration of mononuclear leukocytes, neutrophils, eosinophils, and mast cells. This conclusion is aligned with reports that cellular inflammation is also not a salient feature of human radiation nephropathy. It is reasonable to conclude that mitigation of human renal radiation injury will not be possible by agents acting on inflammatory pathways.

Cellular inflammation found in the present studies was sparse and unimpressive. By way of comparison, cellular inflammation must be present in at least 10% of the parenchyma to be implicated as causing rejection of renal transplants (Loupy et al. 2015). The cellular inflammation in the present studies almost never exceeded 1% of the pa renchymal area.

Immunohistochemical identification of mononuclear leukocytes, such as for the CD45 cell marker, could confirm the present findings. We did not do such staining, but most likely the present conclusions would not be altered because of the very few numbers of visible mononuclear leukocytes.

It is possible that non-cellular inflammation may have a mechanistic role, such as via the presence of one or another cytokine. The presence of IL-22 and its receptor in irradiated kidneys was reported in a recent article. This cytokine appears to have a role in human chronic inflammatory disease, but neither it nor its receptor were present in amounts more than those of non-irradiated NHP (Parker et al. 2019). It remains possible that inflammatory cytokines such as IL-1 or IL-6 could play a mechanistic role, but the cells that would execute their roles were not found in the present studies.

Other organ injuries by irradiation may well have an inflammatory component, such as the alveolar macrophages that are found in radiation pneumonitis. This is relevant to their treatment and mitigation. Indeed, glucocorticosteroids are acknowledged treatments for radiation pneumonitis (Hanania et al. 2019), but those medications do not appear to have a therapeutic role for radiation nephropathy (Berdjis 1960; Caldwell, 1971).

The lack of cellular inflammation is underlined by its lack in the groups of NHP that were treated with leukocyte growth factors. These did not show differences in renal parenchymal cellular inflammation despite the use of leukocyte growth factors, which might be expected to have a pro-inflammatory effect. In the present NHP studies, dexamethasone was used to treat presumed radiation pneumonitis. This was based on indicators including respiratory rate (MacVittie et al. 2012). It is possible that the use of dexamethasone could exert an anti-inflammatory effect, even a lympholytic effect, thus causing the low mononuclear cell counts that were found. However, dexamethasone was never used before d 55 after PBI/BM5 or PBI/BM2.5 in these studies. Its use cannot explain the consistently low mononuclear cell counts in the renal cortex.

Other mechanisms of injury have been proposed for renal radiation injury, such as oxidative stress (Zhao and Robbins 2009). This physico-chemical mechanism is well accepted for the initial prompt effects of radiation injury, but evidence for chronic oxidative stress is limited (Cohen and Cohen 2013). Moreover, antioxidant agents had no mitigating benefit in a rodent model of radiation nephropathy (Cohen et al. 2009). We believe that showing negative results is helpful to justify the investigation of other mechanistic pathways. Neither oxidative stress nor inflammation appear to be significant mechanistic pathways for radiation nephropathy.