The Effects of Thermal Preconditioning on Oncogenic and Intraspinal Cord Growth Features of Human Glioma Cells

Abstract

The adult rodent spinal cord presents an inhibitory environment for donor cell survival, impeding efficiency for xenograft-based modeling of gliomas. We postulated that mild thermal preconditioning may influence the fate of the implanted tumor cells. To test this hypothesis, high-grade human astrocytoma G55 and U87 cells were cultured under 37°C and 38.5°C to mimic regular experimental or core body temperatures of rodents, respectively. In vitro, the 38.5°C-conditioned cells, relative to 37°C, grew slightly faster. Compared to U87 cells, G55 cells demonstrated a greater response to the temperature difference. Hyperthermal culture markedly increased production of Hsp27 in most G55 cells, but only promoted transient expression of cancer stem cell marker CD133 in a small cell subpopulation. We subsequently transplanted G55 cells following 37°C or 38.5°C culture into the C2 or T10 spinal cord of adult female immunodeficient rats (3 rats/each locus/per temperature; total: 12 rats). Systematic analyses revealed that 38.5°C-preconditioned G55 cells grew more malignantly at either C2 or T10 as determined by tumor size, outgrowth profile, resistance to bolus intratumor administration of 5-fluorouracil (0.1 μmol), and posttumor survival (p < 0.05; n = 6/group). Therefore, thermal preconditioning of glioma cells may be an effective way to influence the in vitro and in vivo oncological contour of glioma cells. Future studies are needed for assessing the potential oncogenic modifying effect of hyperthermia regimens on glioma cells.

Keywords

Spinal cord Spinal cord glioma Intramedullary tumor Glioma Glioblastoma Tumor stem cell Thermal stress Preconditioning Heat shock protein (HSP)Cell doubling time

Introduction

Glioblastoma multiforme (GBM) is the most malignant type of glioma (grade IV classified by the World Health Organization) in the central nervous system (CNS). Despite the fact that primary tumors in the spinal cord, spinal nerve roots, and dura are rare compared to intracranial neoplasms, there is a continued increase in clinical incidences (13). Most clinical studies reported that the survival rate for intramedullary spinal cord gliomas (ISCGs) is less than 24 months under maximal therapeutic interventions (22,32). Since effective treatment remains an unmet medical demand due to the poor response to conventional surgical, pharmacological, and radiation management approaches, it is essential to devise effective and efficient modeling systems to investigate pathophysiological mechanisms and develop therapies for ISCGs (29). However, to date, there have been only a few of the experimental studies aiming to investigate spinal cord gliomas. The reality is partly caused by a generally inhibitory environment of the adult mammalian spinal cord, which limits donor cell engraftment (7,14,29). Therefore, the capability of manipulating intraspinal cord engraftment of donor cells is a prerequisite for establishing xenograft-based ISCG models in mammalian species (14,18,29). Determination of proper in vitro culturing regimens (i.e., preconditioning) is an efficacious strategy to prime donor cells (9), and temperature plays a pivotal role in tailoring donor fate in vivo (28). It was reported that cardiac myoblasts with 39.0°C preconditioning were able to increase resistance to oxidative stress postimplantation (36). Because the core body temperature of the rat is normally between 36.5°C and 38.5°C (35,37), we decided to test whether preconditioning human GBM cell lines under a mild thermal environment of either 37.0°C or 38.5°C would affect the intraspinal cord oncological profile of glioma cells by assessing their proliferation and oncological features in vitro as well as growth potential and sensitivity to bolus treatment of 5-fluorouracil (5-FU), a prototype oncolytic drug in vivo.

Materials and Methods

Cell Culture

Human glioblastoma cell lines investigated were U87MG (ATCC® HTB-14^™; WHO grade IV astrocytoma; ATCC, Manassas, VA, USA) and G55. G55 is a human glioblastoma cell line that was passaged in vivo through nude mice and reestablished as a stable xenograft cell line. Dr. E. Y. Snyder of Sanford Burnham Prebys Medical Discovery Institute provided G55 cells that were initially donated by C. David James (Department of Neurological Surgery, University of California, San Francisco, CA, USA) (15,42). G55 cells, showing comparable growth features in vitro as U87 cells, possess signature genes and a vascular endothelial growth factor (VEGF) profile similar to grades III–IV astrocytoma cells (19,30,42); however, because of their in vivo passaging-based establishment, G55 cells maintain more of the characteristics of primary human glioblastoma cells (15). G55 and U87MG cells were maintained with Dulbecco's modified Eagle's medium (DMEM) (Life Technologies, Grand Island, NY, USA) supplemented with 10% fetal bovine serum (FBS) (Atlanta Biologicals, Flowery Branch, GA, USA) and 1% penicillin/streptomycin solution (Life Technologies) in a 37°C and 5% CO₂ incubator. Cells were regularly split with application of 0.25% trypsin (Life Technologies) when they reached ~80% confluency. Because of the quality control records of the providers in maintaining these cell lines, no additional authentication (e.g., DNA fingerprinting) was done for U87 or G55 cells in the present study.

Preconditioning GBM Cells Under 38.5°C Setting

When G55 and U87 cells cultured at 37°C and 5% CO₂ reached 40%–50% confluency, they were replaced into an incubator with an initial setting of 38°C and 5% CO₂ and kept there for 24 h; afterward the incubator's ambient temperature was increased to 38.5°C. When the growing cells reached ~70% confluency (i.e., P0: passage 0), they were split and passaged at a ratio of 1:5. Using the same cell propagation method, the G55 and U87 lines were maintained for two additional passages (i.e., P1 and P2). The total of three passages took about ~9 days for each group.

Calculation of Cell Doubling Time (CDT)

Each passage of cells under 38.5°C preconditioning and their 37.0°C counterparts were detached with trypsin (Life Technologies) and prepared into single-cell suspensions. Cell numbers were counted by a hemocytometer with trypan blue (Sigma-Aldrich, St. Louis, MO, USA) stain being used as the exclusion standard for viability control. CDT was calculated on the basis of the following formula (ATCC):

CDT = T * In 2 / In (N_{T} / N_{0})

N₀ = the number of cells at the beginning of the incubation time

N_T = the number of cells at the end of the incubation duration (i.e., _T)

T = the given duration (unit: hour, for the current study)

Note: T, the incubation duration, could be in any time unit.

Evaluation of Hsp27 Expression

Immunocytochemical stain (ICC) was used to evaluate the expression level of heat shock protein 27 kDa (Hsp27) in G55 cells. Briefly, P0 to P2 glioma cell groups, after 37.0°C or 38.5°C preconditioning, were seeded onto glass coverslips. When the confluency of the cells approached 80%, they were fixed with 2% paraformaldehyde (PFA; Sigma-Aldrich). For the ICC, cell-seeded coverslips were washed with phosphate-buffered saline (PBS) containing Triton X-100 (0.3%; Sigma-Aldrich) and incubated in 5% (v/v) normal donkey serum (EMD Millipore, Billerica, MA, USA) for 30 min. The incubation for the primary antibody against Hsp27 (Catalog No. ADI-SAP-800; Enzo Life Sciences, Farmingdale, NY, USA) took place under 4°C overnight (i.e., ~12 h). Afterward, the specific secondary antibody (Alexa Fluor® 488 AffiniPure Goat Anti-Mouse IgG; Jackson ImmunoResearch Laboratories Inc., West Grove, PA, USA) was applied per specifications provided by the manufacturer. Cells were covered by using mounting medium containing 4′,6-diamidino-2-phenylindole (DAPI; Vector Laboratories, Burlingame, CA, USA) for fluorescence microscopic analysis (Axiovert200; Carl Zeiss, San Diego, CA, USA).

Evaluation of CD133 Expression and Caspase 3 Activation

Briefly, P0 and P2 G55 cells, following 37.0°C or 38.5°C conditioning, were seeded on glass coverslips with an initial density of 7,500 cells/coverslip housed in 12-well culture plates (BD Falcon; BD Biosciences, San Jose, CA, USA). After 72 h, the cells reached 70%–80% confluency before they were fixed with 2% PFA. Using the same preparation procedures, coincubation with primary antibodies against CD133 (MAB4399; EMD Millipore), a cancer stem cell (CSC) marker, and cleaved caspase 3 (antibody #9661; Cell Signaling Technology Inc., Danvers, MA, USA), a cell apoptotic marker, was carried out at 4°C overnight. Corresponding secondary antibodies against mouse and rabbit IgG (Jackson ImmunoResearch Laboratories Inc.), respectively, were applied afterward per specifications provided by the manufacturers. Mounting medium containing DAPI (Vector Laboratories) was used for coverslipping prior to laser confocal fluorescence microscopic analysis (Zeiss LSM1 confocal microscope and Zeiss Zen 2011 software; Carl-Zeiss Microimaging Inc., Jena, Germany).

Quantification of Fluorescence Immunoreactivity of G55 Cells

The percentages of immunopositive cells were quantified by dividing the total positive cell number by the total number of DAPI-labeled nuclei under each randomized 0.25× 0.25-mm² microscopic field. For ICC of Hsp27, three fields for each coverslip and three coverslips per culture condition were quantified (total: n = 9). For quantifying CD133 and cleaved caspase 3 immunopositive cells, five fields for each coverslip and three coverslips per culture condition were evaluated (total: n = 15). Therefore, the estimated total immunopositive cell number per coverslip = the total number of DAPI-labeled nuclei per coverslip × the % of immunopositive cells averaged from the three or five fields randomly sampled in each coverslip.

G55 Xenograft in the Rat Spinal Cord

Surgery

Female immunodeficient rats (8- to 9-week-old RNU; 175–190 g of body weight; Charles River Laboratories, Wilmington, MA, USA) were anesthetized with ketamine hydrochloride (75 mg/kg) and xylazine (10 mg/kg; both were purchased from Patterson Veterinary Supply Inc., Devens, MA, USA) via intraperitoneal (IP) injections. Rats were placed on a heated sterile surgical plate and prepared for surgery. The C2 or T10 spinous process was identified followed by a longitudinal incision made over it using a #10 surgical scalpel (Aspen Surgical, Caledonia, MI, USA). After tissue dissection, laminectomy was made between C2–C3 and T10–T11 for each rat to expose the dorsal surface of the spinal cord.

Tumor Cell Implantation

Using a #11 sharp tip surgical blade (Aspen Surgical), the dura was opened to access the spinal cord, and a 26-gauge needle connected to a Hamilton^™ microsyringe (Hamilton, Reno, NV, USA) was inserted into the dorsal C2–C3 or T10–T11 spinal cord for a 2-mm dorsoventral penetration. The needle was then retracted 0.5 mm prior to microinjection of the P2 G55 cells (10⁴ cells/3 μl of PBS) preconditioned under either a 37°C or a 38.5°C culture setting. The needle was kept inside the spinal cord for an additional 5 min before removal. After muscle and soft tissue closure with surgical sutures (Ethicon 4-0, coated Vicryl suture, undyed braided; Johnson & Johnson, New Brunswick, NJ, USA), the skin incision was stapled (Medline Industries Inc., Mundelein, IL, USA). The rats received the standard postoperation care (29). All experimental procedures were reviewed and approved by the Institutional Animal Care and Use Committee (IACUC) of the Brigham and Women's Hospital and Harvard Medical School.

Evaluation of Pathophysiological Signs

Rats were monitored daily for general physical conditions including body weight, grooming frequency, hair and skin condition, and facial porphyrin staining. Hindlimb functions were also tested daily and summarized weekly based on the Basso, Bresnahan, and Beattie (BBB) scale as previously reported (5). For assessing the tumor's response to a bolus antitumor drug treatment (see below), rats were scheduled to receive a bolus administration of 5-FU (Sigma-Aldrich) when their BBB scores fell to ≤9 unilaterally (i.e., failure to carry out consistent weight-bearing locomotion), a physical sign indicating that they might not be able to fully carry out self-care and should be euthanized (29).

Intratumor Administration of 5-FU

When the BBB score of the hindlimb dropped to ≤9 resulting from ISCG growth, rats were anesthetized and prepared for surgery as described above. The C2–C3 or T10–T11 laminectomy site was reopened in order to examine and access the tumor mass. A bolus microinjection of 10 μl of 5-FU (10 mM; dose: 0.1 μmol) was done over 5 min via a needle inserted close to the center of the tumor mass. The needle was kept inside the tumor for another 2 min before removal. Rats were evaluated for general physical condition and hindlimb BBB score for another 72 h before they were euthanatized by deep anesthesia with IP injection of 90 mg/kg ketamine hydrochloride and 15 mg/kg xylazine (Patterson Veterinary) before systemic perfusion with 4% PFA. The spinal cord and the brain, together with other internal organs, were collected for histopathological and ICC analyses.

Histopathological and ICC Analyses

The postfixed spinal cord tissue was embedded in optimum cutting temperature (OCT) compound (Sakura Finetek USA, Torrance, CA, USA) and cryosectioned transversely at 20-μm thickness. Serial sections (one of every 100- or 500-μm tissue) of the 1.0-cm spinal cord centered at the tumor epicenter were chosen for hematoxylin and eosin (H&E; Sigma-Aldrich) staining for general histopathological analysis of tumor growth. In addition, one cross-section out of every 100 μm of tumor epicenter tissue was selected from each spinal cord for ICC evaluation. Histopathological data of tumor volume were analyzed by creating a computerized three-dimensional (3D) reconstruction of the tumor mass based on serial transverse pathological slices stained with H&E, as described previously (29).

Statistical Analysis

Experimental data are expressed as mean ± SEM. Statistical significance was defined at the p < 0.05 level. Multigroup data were evaluated statistically by using one-way analysis of variance (ANOVA) followed by Tukey's honest significant difference (HSD) post hoc test. Outcomes of two-group studies were compared by paired or unpaired Student's t-tests. We used SPSS (SPSS 13.0; IBM, Armonk, NY, USA) for all data processing.

Results

Mild Thermal Preconditioning Promoted Growth of Human GBM Cells

Under regular tissue culture conditions (i.e., 37.0°C and 5% CO₂), U87 cells and G55 cells had CDTs of 21.55 ± 0.13 h and 19.31 ± 0.11 h, respectively (n = 9/group) (Fig. 1). For the purpose of avoiding excessive stress due to sharply increased temperatures, the two cell lines were initially cultured at 37°C and then underwent a two-step protocol of thermal increment; they were first kept under 38.0°C for a 24-h adaptation period before being exposed to 38.5°C for the continued culture. This design enabled the cells to have uninterrupted growth under different thermal settings without encountering any discernible levels of cell death. Under this protocol, P0 G55 and U87 cells manifested CDTs of 18.80 ± 0.44 h and 20.90 ± 0.42 h, respectively (n = 9/group). With longer incubation times, P1 and P2 G55 cells showed further augmented responses to exposure to a 38.5°C ambient condition. Specifically, the CDT of P2 G55 cells was further shortened to 17.40 ± 0.37 h, which was significantly shorter than that of P2 U87 cells (i.e., 20.40 ± 0.32 h; n = 9/group, p < 0.05, one-way ANOVA with Tukey's HSD test) (Fig. 1). However, the overall CDT differences between the two temperature settings were limited. For example, P2 G55 CDT shortened about 1.91 h (i.e., <10%) on average under 38.5°C compared to 37.0°C. Correspondingly, there was only a 16.60 ± 0.21% increase in the total cell number per coverslip for 38.5°C-cultured P2 G55 cells versus 37.0°C-cultured P0 G55 cells (n = 9/group). The result indicated that the designed thermal conditioning had no dramatic impact on glioma cell proliferation. In either group, no noticeable levels of cell death were present, suggesting that the thermally escalated cell proliferation was not caused by elimination of temperature-sensitive cell subpopulations. We therefore decided to focus on G55 cells for all of the subsequent assays (specifically on biological and oncological features of G55 cells) (26,27).

Figure 1.

Promotion of glioma cell growth in vitro by mild hyperthermia. U87 and G55 glioblastoma cells showed significantly augmented cell proliferation rates when cultured in a 38.5°C ambient condition for one to two passages (i.e., P1 and P2). Group average cell doubling time (CDT), relative to P0 baseline values, was significantly shortened in U87 [e.g., from P0 CDT of 21.55 ± 0.13 h to P2 CDT of 20.40 ± 0.32 h, n = 9, p < 0.05, one-way analysis of variance (ANOVA) with Tukey's honest significant difference (HSD) test] and in G55 cells (e.g., from P0 CDT of 19.31 ± 0.11 to P2 CDT of 17.40 ± 0.37 h, n = 9, p < 0.05, one-way ANOVA with Tukey's HSD test).

Increased Hsp27 Expression in GBM Cells Following 38.5°C Preconditioning

In order to determine whether there were any cytological changes resulting from the thermal stress, which might underlie the 38.5°C exposure-triggered more robust cell division, ICC evaluation of Hsp27 expression was performed. Hsp27 is a member of a small HSP family with well-characterized functions of protein chaperone activity, thermotolerance, inhibition of apoptosis, and regulation of either normal or malignant cell growth (2,31). Under the 37°C culture condition, only a fraction of P1 G55 cells synthesized Hsp27 at detectible levels (i.e., 7.41 ± 2.23% of the cells counted, n = 9/group) (Fig. 2). However, when P1 G55 cells were cultured under 38.5°C, Hsp27 expression level was significantly elevated (i.e., 72.52 ± 1.77% of the total cell growth field; n = 9, p < 0.05, one-way ANOVA with Tukey's HSD test) (Fig. 2). Interestingly, the increase in Hsp27 production in G55 cells following thermal exposure was partially a time-dependent event. For example, after ~9 days of continued 38.5°C culture, the percentage of Hsp27 immunopositive cells decreased to 60.33 ± 2.28% in P2 G55 cells, though it remained significantly higher than that of 37°C P1 G55 cells (n = 9, p < 0.05, one-way ANOVA with Tukey's HSD test) (Fig. 2), suggesting that the cells were progressively acclimated to the 38.5°C culture condition by adjusting their essential biological processes such as growth rate (Fig. 1), resistance to oxidative insults, and thermotolerance (e.g., upregulation of Hsp27) (2,26,27,31).

Figure 2.

Upregulation of Hsp27 expression in glioma cells following 38.5°C preconditioning. Under 37°C culture, about 7.41 ± 2.23% of G55 cells showed immunoreactivity for heat shock protein 27 kDa (Hsp27) (A). By contrast, when the cells were cultured at 38.5°C (P0), the percentage of cells immunopositive for Hsp27 increased significantly to 72.52 ± 1.77% (n = 9, p < 0.05, one-way ANOVA with Tukey's HSD test) (B). However, when G55 cells were continuously cultured at 38.5°C for three passages (P2: ~9 days), the percentage of Hsp27 immunopositive cells decreased to 60.33 ± 2.28% (n = 9, p < 0.05, one-way ANOVA with Tukey's HSD test) (C). A summary of the percentage of Hsp27 immunopositive cells in different groups is presented in (D). Insets: Typical density of DAPI-labeled nuclei of ~80% confluent cells that were cultured under different temperatures. Percentage of Hsp27 immunocytochemical stain (ICC)-positive cells was quantified via dividing the total Hsp27 immunopositive cell number by the total number of DAPI-labeled nuclei under each randomly sampled area (see Materials and Methods for details). Scale bar: 10 μm. ∗p < 0.05 compared to cells in the 37°C culture condition; #p < 0.05 compared to P0 cells in the 38.5°C culture condition.

Thermal Culture Transiently Upregulated Cancer Stem Cell Marker CD133 in a Subpopulation of G55 Cells That Were Apoptosis Resistant

We next assessed the effect of thermal conditioning on the oncological feature of G55 cells by evaluating levels of CSC marker CD133 expression (10,11,43). In a 37°C setting, 2.7 ± 0.16% of G55 cells expressed CD133. However, CD133 expression percentage of 38.5°C-cultured P0 G55 cells increased to 6.48 ± 0.36, but subsequently decreased to 3.96 ± 0.13 in P2 cells, both being significantly higher than at the 37°C setting (n = 15/group, p < 0.05, ANOVA with Tukey's HSD test) (Fig. 3A–D). We then verified whether CD133-expressing G55 cells were more apoptosis resistant, a key oncological feature of CSCs (4,10,11,43). Following 38.5°C of culture, there was a significantly increased rate of apoptosis in G55 P0 cells, but the degree of increase was very limited (i.e., 2.34 ± 0.13% vs. 1.64 ± 0.16% of 37.0°C; n = 15/group, p < 0.05, ANOVA with Tukey's HSD test) (Fig. 3E–H) as detected by nuclear presence of cleaved caspase 3. It has been known that hyperthermotherapy might enhance the efficacy of chemotherapy on clinical gliomas by increasing tumor cell apoptosis (16). The apoptotic cell percentage decreased to 1.56 ± 0.10% in 38.5°C P2 cells, which was insignificant compared to 37°C control cells. In all of the settings, immunoreactivity to cleaved caspase 3 was only observed in G55 cells that were negative for CD133 expression (Fig. 3E–G; note that A–D, respectively, showed the same imaging fields in E–G).

Figure 3.

Culture at 38.5°C transiently elevated CD133 expression in a G55 subpopulation but did not trigger caspase 3 cleavage in those cells. In a 37°C setting, 2.7 ± 0.16% of G55 cells expressed CD133. However, CD133 expression percentage of 38.5°C-cultured P0 G55 cells increased to 6.48 ± 0.36, but subsequently decreased to 3.96 ± 0.13 in P2 cells, showing a transient pattern, though the latter remaining significantly higher than control levels (n = 15/group, p < 0.05, ANOVA with Tukey's HSD test) (A–D). Following 38.5°C of culture, there was a significantly increased rate of apoptosis in G55 P0 cells, but the degree of increase was very limited (i.e., 2.34 ± 0.13% vs. 1.64 ± 0.16% of 37.0°C; n = 15/group, p < 0.05, ANOVA with Tukey's HSD test) (E–H) as detected by nuclear presence of cleaved caspase 3 (see details in the orthoslice images in the larger inset). The apoptotic cell percentage decreased to 1.56 ± 0.10% in 38.5°C P2 cells, which was insignificant compared to that in 37°C control cells (1.64 ± 0.16%). In all the settings, immunoreactivity to cleaved caspase 3 was only observed in G55 cells that were negative for CD133 expression (E–G; note: A–D, respectively, show the same imaging fields of E–G). The total cell number estimates are presented in (D) and (H). Insets: Typical density of DAPI-labeled nuclei of ~70–80% confluent cells that were cultured at different temperatures.

In Vivo Evaluation of 38.5°C-Preconditioned G55 Cells in the Rat Spinal Cord for Their Oncological Features

As a pathophysiological sign of ISCG growth, rats showed gradually heightened impairment of hindlimb locomotion and eventually lost the ability to perform body weight-bearing stepping (i.e., BBB score ≤9) (44). Therefore, per humane standards for animal care, the first daily detection on loss of weight-bearing stepping in the hindlimb was used to determine posttumor survival duration (29). Because there was no significant statistical difference in the survival length between C2 (n = 3 × 2 temperature groups) and T10 (n = 3 × 2 temperature groups) implantation of G55 cells (Table 1), the data of C2 and T10 groups were combined to assess whether there was any effect of 38.5°C or 37°C preconditioning on posttumor survival. Indeed, immunodeficient rats with implantation of 38.5°C-preconditioned G55 cells showed significantly shortened group mean survival time than the group that received tumor cells preconditioned under 37°C (15.33 ± 0.67 days vs. 18.33 ± 0.71 days, n = 6/group, p = 0.012, unpaired Student's t-test). The outcome suggested that relative to 37°C exposure, G55 cells after 38.5°C preconditioning had augmented malignancy in the rat spinal cord, resulting in a more rapid deterioration of neural functions (Fig. 4 and Table 1).

Figure 4.

In vivo oncological profile of G55 after 38.5°C preconditioning in the spinal cord. Implantation of 38.5°C-preconditioned G55 cells resulted in significantly larger outgrowth of the tumor compared with that derived from 37°C-preconditioned G55 at the C2 and T10 spinal cord as measured by the length and width of the glioma mass (A). As a physical sign of tumor progress, rats showed gradual impairment of the hindlimb locomotor function and eventually lost their ability to perform body weight-bearing stepping (i.e., BBB score ≤9). This was the time when intratumor 5-FU (dose: 0.1 μmol) was administered to assess sensitivity to chemotherapy (i.e., oncological aggressiveness). Rats with 38.5°C-preconditioned G55 implants at either C2 or T10 had very poor responses to bolus 5-FU treatment, showing continued deterioration of BBB scores 72 h later; by contrast, rats transplanted with 37°C-cultured G55 cells responded well to the treatment, with their BBB scores being maintained at similar levels showing slight improvement (B). In fact, the group BBB score change (i.e., BBB at the day of 5-FU injection – BBB at 72 h after 5-FU) in average was 3.33 ± 0.42 for the 38.5°C group versus −1.33 ± 0.91 for the 37.0°C group (n = 6/group; p < 0.01, paired Student's t-test) (C; right column). The data suggested that 38.5°C preconditioning made G55 cells more malignant. The notion was validated by rat posttumor survival evaluation. Specifically, rats with 38.5°C-preconditioned G55 cells in the C2 or T10 spinal cord showed a group mean survival of 15.33 ± 0.67 days after G55 transplantation, which was significantly shorter than that of rats that received 37°C-preconditioned G55 cells (18.33 ± 0.71 days, n = 6/group, p = 0.012, unpaired Student's t-test) (C; left column). ∗p < 0.05 when compared to rats receiving 37°C-preconditioned G55 cells.

Table 1.

In Vivo Profiles of G55 Growth and Response to 5-FU Treatment After Intraspinal Cord Implantation

ID	Thermal Condition/Tumor Site	Tumor Length (cm)	Tumor Width (cm)	Tumor Outgrowth	The First Daily BBB Score to ≤9 (Unit: Score)	BBB Scale at 72 h After 5-FU Injection (Unit: Score)	Survival Duration (First Daily BBB ≤9) (Unit: Day)
1	38.5°C/G55, C2	1.3	0.7	Dorsal dura disruption	9	6	15
7	38.5°C/G55, C2	0.7	0.5	Dorsal dura disruption	8	6	13
9	38.5°C/G55, C2	0.8	0.6	Dorsal dura disruption	9	5	16
8	38.5°C/G55, T10	0.6	0.5	Dorsal dura disruption	6	3	15
10	38.5°C/G55, T10	0.6	0.3	Dorsal disruption	8	5	17
12	38.5°C/G55, T10	0.8	0.3	Inside spinal cord	8	3	14
2	37°C/G55, C2	0.6	0.3	Inside spinal cord	9	8	18
5	37°C/G55, C2	0.5	0.3	Inside spinal cord	9	11	19
11	37°C/G55, C2	0.4	0.3	Inside spinal cord	9	10	16
3	37°C/G55, T10	0.2	0.1	Inside spinal cord	7	9	17
4	37°C/G55, T10	0.3	0.2	Inside spinal cord	9	14	19
6	37°C/G55, T10	0.4	0.3	Inside spinal cord	9	8	21

We next examined if 38.5°C preconditioning also altered the oncological resilience of G55 cells in vivo. This was done by evaluating the resistance degree of intraspinal cord G55 cells to a bolus intratumor administration of 5-FU given to the rat on the day when its BBB score dropped to ≤9 (i.e., longevity endpoint; see Materials and Methods for more details). The drug 5-FU inhibits tumor growth by disrupting DNA elongation to trigger apoptosis (29). At 72 h after 5-FU administration, rats with 38.5°C-preconditioned G55 glioma demonstrated significantly poorer response to 5-FU chemotherapy compared to the 37°C group, as determined by further worsened mean BBB score (i.e., score difference between the first day of BBB ≤9 and that observed at 72 h after 5-FU treatment). Rats with 38.5°C-preconditioned G55 glioma had a group mean individual BBB score difference of 3.33 ± 0.42 (i.e., more than a 3-point loss in average) versus −1.33 ± 0.91 (i.e., >1 point improvement) of the 37°C group (n = 6/group, p = 0.001, paired Student's t-test). In fact, none of the rats of the 38.5°C-preconditioned G55 group showed motor function improvement following 5-FU administration; moreover, five out of the six rats had BBB scores that decreased >3 points. By sharp contrast, four out of the six rats that received 37°C-preconditioned G55 cells manifested motor functional improvement following bolus 5-FU treatment, showing the best BBB score improvement by 5 points (i.e., from 9 to 14) (Table 1 and Fig. 4). The data indicated that mild thermal preconditioning could markedly influence the oncogenic features of human high-grade glioma cells in the adult mammalian spinal cord (28).

Pathological Features of Intraspinal Cord Growth of 38.5°C-Preconditioned G55 cells

There were strikingly different gross pathological profiles of gliomas derived from implantation of 38.5°C-preconditioned G55 cells relative to those derived from 37°C-cultured tumor cells (Fig. 5). Among the six spinal cords that were injected with 38.5°C-preconditioned G55 cells, five exhibited obvious tumor outgrowth disrupting the dorsal surface of the spinal cord (Fig. 5, upper panel), whereas none of the glioma masses derived from G55 cells after 37°C preconditioning grew above the dorsal surface, leaving the spinal cord with a relatively smooth and largely intact dura membrane (Fig. 5, lower panel). Furthermore, tumor size in rats that received 37°C-preconditioned G55 cells, on average, appeared smaller (Fig. 5 and Table 1). H&E staining histochemically validated the aforementioned conclusions (Fig. 5). Therefore, 38.5°C preconditioning resulted in more robust outgrowth of G55 gliomas and higher resistance to chemotherapeutic effects that impede cell proliferation and viability (e.g., intratumor delivery of 5-FU) (Table 1).

Figure 5.

Pathological features of thermal preconditioned G55 glioma in the spinal cord. Five out of the six spinal cords showed dorsal surface disruption by the outgrowth of glioma mass derived from the implanted G55 cells following 38.5°C preconditioning (upper row). By contrast, none of the tumors derived from 37°C-preconditioned G55 cells grew above the dorsal surface in the six spinal cords (lower row). Right: hematoxylin and eosin (H&E) stain demonstrated representative pathological scale of the intramedullary growth of the glioma mass for each group.

Discussion

The adult mammalian spinal cord, in general, presents a dormant and inhibitory environment for endogenous stem cells and donor cell engraftment. The feature sets up barriers not only for regenerative biology-based neural repair approaches but also for efforts to establish xenograft models of spinal cord disorders such as ISCG. In the current study, we observed that culture temperature differences exerted greater effect on G55 than U87 human glioma cells (e.g., CDT reduction). G55 cells preconditioned in 38.5°C, relative to 37°C, showed limited increases in proliferation and apoptosis rates; however, their Hsp27 and CD133 expressions were markedly augmented. Compared to sustainable levels of Hsp27 increases, elevated expression of CD133 was transient in a G55 subpopulation. Importantly, 38.5°C culture enabled G55 cells to grow more robustly at either the C2 or T10 spinal cord as determined by tumor size and growth profile, or resistance affecting anticancer efficacy of a bolus intratumor administration of 5-FU. We conclude that mild thermal preconditioning effectively influences the oncological features of human glioma cells in vitro and in vivo. Our data suggest that future studies should determine potential oncogenic modifying effects of hyperthermia regimens on glioma cells in order to devise efficacious ways to manipulate glioma cell fate.

Clinical incidence of ISCG is generally rare relative to that of brain glioblastomas. However, it could carry an even poorer prognosis for many patients. Moreover, no targeted treatment has been established as either primary or adjuvant treatments for clinical spinal gliomas. This reality is partly caused by difficulties encountered in endeavors trying to establish clinically relevant ISCG models (29). The temperature of the CNS of mammalian species, though conventionally ranged under “core body temperature,” could even be ~1°C higher than that of most of the other internal organs due to its vigorous metabolic activity (1). Published work demonstrated that besides directly modifying interactions with neighboring cells and/or other environmental elements (23,33), the heat shock pretreatment could be a practical method to influence cell fate in order to improve the success of the donor cell engraftment (6). We thereby designed and tested whether preculturing human high-grade glioma cells under a 37°C or 38.5°C ambient condition could impact the oncological features of U87 and G55, two glioma cell lines that may have different genetic and epigenetic profiles (42). It has been reported that moderate heat shock exposure (e.g., 43°C for 1 or 2 h), clinically applied to tumor–brain border zone, caused cancer cell acute autophagy and cell cycle arrest, but did not induce apparent tumor cell apoptosis (20,45). However, to date, very limited attention has been given to the effect of even milder temperature elevation (e.g., 38.5°C used in the current study) on the oncological features of human glioma cells. Our data show that whereas the two-step incremental exposure from 38°C to 38.5°C culture slightly reduced CDT in U87 and G55 cells, it discernibly impacted the expression paradigms of Hsp27 and CSC marker CD133 in G55 cells. Furthermore, our mild hyperthermia regimen heightened engraftment, growth, and viability resilience of G55 cells in the adult rat spinal cord. At least part of the preconditioning-triggered changes could be underlined by 38.5°C-induced upregulation of Hsp27. As a chaperone protein responding to heat stress, Hsp27 is one of the most widely studied heat shock proteins in oncology since it plays a major role in inhibiting extrinsic and intrinsic cell death signaling pathways to reduce apoptosis (3,8). Hsp27 is also related to the state of tumor growth and to the aggressive state of tumor cells (12,31,40,41). Pathological overexpression of Hsp27 has been reported in a wide spectrum of malignancies, which was associated with poor prognosis in different types of tumors, including high-grade invasive gliomas. Correlation analysis on Hsp27 expression and tumor growth rate of astrocytomas revealed that abnormal levels of Hsp27 were likely to promote tumor growth (17). It has also been reported that the majority of cells in a given GBM mass could be generated by a very small fraction of self-renewing, multipotent tumor-initiating cells/CSC that may be accountable for tumor growth, recurrence, and resistance to chemo- and radiotherapies (4,10, 24,34). Indeed, in addition to sustainable increases in Hsp27, 38.5°C-cultured G55 cells transiently augmented CD133 expression in a small subpopulation. CD133 is a glycoprotein that in humans is encoded by the prominin-1 (PROM1) gene that has been considered as a representative marker of CSCs (4,10,43). Thus, our results that rats receiving 38.5°C-preconditioned G55 cells demonstrated more aggressively developing tumors that were less sensitive to 5-FU treatment suggest that milder heat stress exposures may induce transiently increased expression of CD133 in a fraction of glioma cells that may be more resilient for survival and proliferation (e.g., as per our in vitro data, none of the CD133-positive G55 cells exhibited cleaved caspase 3, an apoptotic marker following bolus 5-FU treatment). Lately, analyses based on bench data and mathematical modeling indicated a possibility that stem cell-like tumor initiation cells may not be a fixed population within any growing tumor. Instead, expressions of CSC markers may likely be transient oncological events occurring in a selected subpopulation of cancer cells when induced by cell growth dynamics as well as environmental, epigenetic, and genetic impacts (25,26). We are currently exploring whether mild hyperthermiastressed glioma cells also have constitutively activated Hsp27 in CD133⁺ cells as reported before for CSCs under hypoxia and serum depletion to inhibit caspase activation (21). Although hyperthermia therapy has been conventionally reasoned as a putative treatment for different types of malignant tumors including GBM, data from large, controlled studies remain lacking (39). Conversely, fever may play a detrimental role by increasing malignant severity of glioma cells since it is one of the most common signs for the initial diagnosis and in the terminal stage of GBM patients (38). Therefore, systematic studies will be needed to examine potential oncogenic modifying or enhancing effects of hyperthermia regimens on glioma and other types of cancer cells by applying stem cell biology-oriented designs (24,25).

We recently reported that genetically engineered human neural stem cells (hNSCs) plus prodrugs effectively treated glioblastoma in the first rat model of ISCG manifesting both somatomotor and autonomic abnormalities (29). The present study has, in addition, established an effective approach of thermal preconditioning of human high-grade glioma cells for altering their oncological features in vitro and in vivo. Using the mild thermal exposure tactic, we have developed a protocol of manipulating growth dynamics of human glioma cells in the adult rat spinal cord. Taken together, our work provides new experimental systems for investigating oncogenic mechanisms and screening for potential oncolytic therapeutics to treat spinal cord malignant tumors. The findings may additionally facilitate future investigations of the impact of hyperthermia on malignant tumor cells and therapeutic discoveries for currently intractable metastatic diseases.

Footnotes

Acknowledgments

This work was supported by Teng Lab Research Fund, CASIS-NASA, and SCIRP-DoD grants to Y.D.T., and a grant to Y.D.T. and R.D.Z. from the Cele H. and William B. Rubin Family Fund Inc. for the Gordon Program in Clinical Paralysis Research. We thank Dr. E. Y. Snyder of SBPDC for providing G55 cells. X.Z., I.B.H., M.A., Z.A., J.E.A., and Y.D.T. undertook major experimentation plus data processing. X.Z., I.B.H., and M.A. contributed to the study design refinement. R.D.Z. and J.H.C. provided partial grant support and advice for the study design and data analysis. X.Z. and Y.D.T. drafted the manuscript. Y.D.T. conceived and designed the study, provided funding, supervised all experimentation, performed data analyses, and wrote and finalized the manuscript. The authors declare no conflicts of interest.

References

Abrams

; Hammel

H. T.

Hypothalamic temperature in unanesthetized albino rats during feeding and sleeping. Am. J. Physiol. 206: 641–646; 1964.

Arrigo

A. P.

In search of the molecular mechanism by which small stress proteins counteract apoptosis during cellular differentiation. J. Cell. Biochem. 94(2): 241–246; 2005.

Arya

; Mallik

; Lakhotia

S. C.

Heat shock genes—Integrating cell survival and death. J. Biosci. 32(3): 595–610; 2007.

Bao

; Wu

; McLendon

R. E.

; Hao

; Shi

; Hjelmeland

A. B.

; Dewhirst

M. W.

; Bigner

D. D.

; Rich

J. N.

Glioma stem cells promote radioresistance by preferential activation of the DNA damage response. Nature 444(7120): 756–760; 2006.

Basso

D. M.

; Beattie

M. S.

; Bresnahan

J. C.

A sensitive and reliable locomotor rating scale for open field testing in rats. J. Neurotrauma 12(1): 1–21; 1995.

Bouchentouf

; Benabdallah

B. F.

; Tremblay

J. P.

Myoblast survival enhancement and transplantation success improvement by heat-shock treatment in mdx mice. Transplantation 77(9): 1349–1356; 2004.

Caplan

; Pradilla

; Hdeib

; Tyler

B. M.

; Legnani

F. G.

; Bagley

C. A.

; Brem

; Jallo

A novel model of intramedullary spinal cord tumors in rats: Functional progression and histopathological characterization. Neurosurgery 59(1): 193–200; 2006.

Concannon

C. G.

; Gorman

A. M.

; Samali

On the role of Hsp27 in regulating apoptosis. Apoptosis 8(1): 61–70; 2003.

Drela

; Sarnowska

; Siedlecka

; Szablowska-Gadomska

; Wielgos

; Jurga

; Lukomska

; Domanska-Janik

Low oxygen atmosphere facilitates proliferation and maintains undifferentiated state of umbilical cord mesenchymal stem cells in an hypoxia inducible factor-dependent manner. Cytotherapy 16(7): 881–892; 2014.

10.

Fidler

I. J.

The organ microenvironment and cancer metastasis. Differentiation 70(9-10): 498–505; 2002.

11.

Galli

; Binda

; Orfanelli

; Cipelletti

; Gritti

; De Vitis

; Fiocco

; Foroni

; Dimeco

; Vescovi

Isolation and characterization of tumorigenic, stem-like neural precursors from human glioblastoma. Cancer Res. 64(19): 7011–7021; 2004.

12.

Garrido

; Brunet

; Didelot

; Zermati

; Schmitt

; Kroemer

Heat shock proteins 27 and 70: Anti-apoptotic proteins with tumorigenic properties. Cell Cycle 5(22): 2592–2601; 2006.

13.

Hsu

; Quattrone

; Ostrom

; Ryken

T. C.

; Sloan

A. E.

; Barnholtz-Sloan

J. S.

Incidence patterns for primary malignant spinal cord gliomas: A Surveillance, Epidemiology, and End Results (SEER) study. J. Neurosurg. Spine 14(6): 742–747; 2011.

14.

Hsu

; Siu

I. M.

; Pradilla

; Gokaslan

Z. L.

; Jallo

G. I.

; Gallia

G. L.

Animal model of intramedullary spinal cord glioma using human glioblastoma multiforme neurospheres. J. Neurosurg. Spine 16(3): 315–319; 2012.

15.

Ito

; Ohba

; Gaensler

; Ronen

S. M.

; Mukherjee

; Pieper

R. O.

Early Chk1 phosphorylation is driven by temozolomide-induced, DNA double strand break- and mismatch repair-independent DNA damage. PLoS One 8(5): e62351; 2013.

16.

; Yamaguchi

; Zhan

; Higuchi

; Asakura

; Morita

; Orimo

; Hu

Hyperthermotherapy enhances antitumor effect of 5-aminolevulinic acid-mediated sonodynamic therapy with activation of caspase-dependent apoptotic pathway in human glioma. Tumour Biol. 37(8): 10415–10426; 2016.

17.

Khalid

; Tsutsumi

; Yamashita

; Kishikawa

; Yasunaga

; Shibata

Expression of the small heat shock protein (hsp) 27 in human astrocytomas correlates with histologic grades and tumor growth fractions. Cell. Mol. Neurobiol. 15(2): 257–268; 1995.

18.

Kim

; Cooke

M. J.

; Shoichet

M. S.

Creating permissive microenvironments for stem cell transplantation into the central nervous system. Trends Biotechnol. 30(1): 55–63; 2012.

19.

Kim

K. J.

; Li

; Winer

; Armanini

; Gillett

; Phillips

H. S.

; Ferrara

Inhibition of vascular endothelial growth factor-induced angiogenesis suppresses tumor growth in vivo. Nature 362: 841–844; 1993.

20.

Komata

; Kanzawa

; Nashimoto

; Aoki

; Endo

; Nameta

; Takahashi

; Yamamoto

; Kondo

; Tanaka

Mild heat shock induces autophagic growth arrest, but not apoptosis in U251-MG and U87-MG human malignant glioma cells. J. Neurooncol. 68(2): 101–111; 2004.

21.

Lin

S. P.

; Lee

Y. T.

; Wang

J. Y.

; Miller

S. A.

; Chiou

S. H.

; Hung

M. C.

; Hung

S. C.

Survival of cancer stem cells under hypoxia and serum depletion via decrease in PP2A activity and activation of p38-MAPKAPK2-Hsp27. PLoS One 7(11): e49605; 2012.

22.

Matsumoto

; Urasaki

; Soejima

; Nakano

; Yokota

; Nishizawa

Cervical intramedullary glioblastoma: Report of a long-term survival case and a review of the literature. J. UOEH 30(4): 413–420; 2008.

23.

Motaln

; Koren

; Gruden

; Ramšak

Schichor

; Lah

T. T.

Heterogeneous glioblastoma cell cross-talk promotes phenotype alterations and enhanced drug resistance. Oncotarget 6(38): 40998–41017; 2015.

24.

Munoz

J. L.

; Rodriguez-Cruz

; Greco

S. J.

; Ramkissoon

S. H.

; Ligon

K. L.

; Rameshwar

Temozolomide resistance in glioblastoma cells occurs partly through epidermal growth factor receptor-mediated induction of connexin 43. Cell Death Dis. 5: e1145; 2014.

25.

Pillat

M. M.

; Oliveira

M. N.

; Motaln

; Breznik

; Glaser

; Lah

T. T.

; Ulrich

Glioblastoma-mesenchymal stem cell communication modulates expression patterns of kinin receptors: Possible involvement of bradykinin in information flow. Cytometry A 89(4): 365–375; 2015.

26.

Poleszczuk

; Enderling

Cancer stem cell plasticity as tumor growth promoter and catalyst of population collapse. Stem Cells Int. 2016: 3923527; 2016.

27.

Poleszczuk

; Hahnfeldt

; Enderling

Evolution and phenotypic selection of cancer stem cells. PLoS Comput. Biol. 11(3): e1004025; 2015

28.

Rajan

; Eubanks

; Edwards

; Aronovich

; Travan

; Rudek

; Wang

; Lanis

; Kaigler

Optimized cell survival and seeding efficiency for craniofacial tissue engineering using clinical stem cell therapy. Stem Cells Transl. Med. 3(12): 1495–1503; 2014.

29.

Ropper

A. E.

; Zeng

; Haragopal

; Anderson

J. E.

; Aljuboori

; Han

; Abd-El-Barr

; Lee

H. J.

; Sidman

R. L.

; Snyder

E. Y.

, Viapiano

M. S.

; Kim

S. U.

; Chi

J. H.

; Teng

Y. D.

Targeted treatment of experimental spinal cord glioma with dual gene-engineered human neural stem cells. Neurosurgery 79(3): 481–491; 2016.

30.

Rubenstein

J. L.

; Kim

; Ozawa

; Zhang

; Westphal

; Deen

D. F.

; Shuman

M. A.

Anti-VEGF antibody treatment of glioblastoma prolongs survival but results in increased vascular cooption. Neoplasia 2(4): 306–314; 2000.

31.

Samali

; Holmberg

C. I.

; Sistonen

; Orrenius

Thermotolerance and cell death are distinct cellular responses to stress: Dependence on heat shock proteins. FEBS Lett. 461(3): 306–310; 1999.

32.

Samartzis

; Gillis

C. C.

; Shih

; O'Toole

J. E.

; Fessler

R. G.

Intramedullary spinal cord tumors: Part I—Epidemiology, pathophysiology, and diagnosis. Global Spine J. 5(5): 425–435; 2015.

33.

Seidel

; Garvalov

B. K.

; Wirta

; von Stechow

; Schänzer

, Meletis

; Wolter

; Sommerlad

; Henze

A. T.

; Nister

; Reifenberger

; Lundeberg

; Frisen

; Acker

A hypoxic niche regulates glioblastoma stem cells through hypoxia inducible factor 2 alpha. Brain 133(Pt 4): 983–995; 2010.

34.

Singh

S. K.

; Hawkins

; Clarke

I. D.

; Squire

J. A.

; Bayani

; Hide

; Henkelman

R. M.

; Cusimano

M. D.

; Dirks

P. B.

Identification of human brain tumour initiating cells. Nature 432(7015): 396–401; 2004.

35.

Spencer

; Shirer

H. W.

; Yochim

J. M.

Core temperature in the female rat: Effect of pinealectomy or altered lighting. Am. J. Physiol. 231(2): 355–360; 1976.

36.

C. Y.

; Chong

K. Y.

; Chen

; Ryter

; Khardori

; Lai

C. C.

A physiologically relevant hyperthermia selectively activates constitutive hsp70 in H9c2 cardiac myoblasts and confers oxidative protection. J. Mol. Cell. Cardiol. 31(4): 845–855; 1999.

37.

Teng

Y. D.

; Wrathall

J. R.

Evaluation of cardiorespiratory parameters in rats after spinal cord trauma and treatment with NBQX, an antagonist of excitatory amino acid receptors. Neurosci. Lett. 209(1): 5–8; 1996.

38.

Their

; Calabek

; Tinchon

; Grisold

; Oberndorfer

The last 10 days of patients with glioblastoma: Assessment of clinical signs and symptoms as well as treatment. Am. J. Hosp. Palliat. Care 33(10): 985–988; 2016.

39.

Titsworth

W. L.

; Murad

G. J.

; Hoh

B. L.

; Rahman

Fighting fire with fire: The revival of thermotherapy for gliomas. Anticancer Res. 34(2): 565–574; 2014.

40.

Voll

E. A.

; Ogden

I. M.

; Pavese

J. M.

; Huang

; Xu

; Jovanovic

B. D.

; Bergan

R. C.

Heat shock protein 27 regulates human prostate cancer cell motility and metastatic progression. Oncotarget 5(9): 2648–2663; 2014.

41.

Wang

; Chen

; Zhou

; Zhang

HSP27, 70 and 90, anti-apoptotic proteins, in clinical cancer therapy (Review). Int. J. Oncol. 45(1): 18–30; 2014.

42.

Westphal

; Hansel

; Hamel

; Kunzmann

; Holzel

Karyotype analyses of 20 human glioma cell lines. Acta Neurochir. (Wien) 126: 17–26; 1994.

43.

; Sun

; Feng

; Ge

; Xia

Do relevant markers of cancer stem cells CD133 and Nestin indicate a poor prognosis in glioma patients?

A systematic review and meta-analysis. J. Exp. Clin. Cancer Res. 34: 44; 2015.

44.

; Thakor

D. K.

; Han

; Ropper

A. E.

; Haragopal

; Sidman

R. L.

; Zafonte

; Schachter

S. C.

; Teng

Y. D.

Alleviation of chronic pain following rat spinal cord compression injury with multimodal actions of huperzine A. Proc. Natl. Acad. Sci. USA 110(8): E746–E755; 2013.

45.

Zolzer

; Streffer

Quiescence in S-phase and G1 arrest induced by irradiation and/or hyperthermia in six human tumour cell lines of different p53 status. Int. J. Radiat. Biol. 76(5): 717–725; 2000.