Human glioma cells (U87, U373, LN229 and T98G) were cultured in Dulbecco's modified eagle's medium (DMEM), supplemented with 10% fetal calf serum (GIBCO, USA) and 1% Penicillin/Streptomycin (SIGMA ALDRICH, USA), and incubated at 37°C with 5% CO₂. It should be specially noted that the origin of the U87 cell line is unknown, but it is a likely glioblastoma cell line (6). And U373 cells are known to be a U251 derivative (7), obtained from ATCC (HTB-17).

All the study specific tissue samples were collected from the glioma patients treated at the department of neurosurgery, Tangdu hospital, Fourth Military Medical University, P.R. China. These tissue samples comprised of 4 glioma (1G: WHO IV, 2G: WHO III, 3G: WHO III–IV, 4G: WHO IV) + 4 adjacent normal brain tissue specimens, freshly collected from the surgeries during 2015 and 54 glioma only specimens, archived under liquid nitrogen at the pathology department, from the surgeries during 01/2008 to 12/2010. None of the patients had received chemotherapy or radiotherapy before surgery. The fresh specimens were snap-frozen in liquid nitrogen and stored at −80°C, while archived glioma only specimens were fixed in 10% buffered formalin and embedded in paraffin, till further analysis.

The total RNA from human glioma cell lines (U87, U373, LN229 and T98G) was purified using TRIzol reagent. Normal brain tissue RNA was purchased from Clontech. Typically, 1 µg of the total RNA was used to generate cDNA by employing SuperScript^® II RT (TAKARA, JAPAN) with oligo-dT primer. The PCR primers for IRX1 and GAPDH were 5′-TCATTGACCTCAACTACATG-3′ (F) and 5′-TCGCTCCTGGAAGATGGTGAT-3′ (R), and 5′-CTCAGCCTCTTCTCGCAGAT-3′ (F) and 5′-TCTTCCTGGTCCTTGCTGC-3′ (R), respectively (8). Each PCR was performed for 30 thermal cycles and the PCR products were observed by electrophoresis on a 1.5% agarose gel, visualized after staining with ethidium bromide.

The IRX1 protein expression in the glioma condition was assessed using the fresh glioma tissues, along with the adjacent normal brain tissue. The total protein from fresh tissues, separated into glioma tissue and adjacent normal tissue, were extracted via mammalian protein extraction reagent (Pierce, Appleton, WI, USA) supplemented with protease inhibitors cocktail (Sigma, USA). The total protein concentrations were determined via BCA protein assay kit (Boster Systems, Inc., Pleasanton, CA, USA). The proteins were separated by 10% SDS-PAGE (Genshare biological, shaanxi, China) and then transferred to PVDF membranes (Millipore, Boston, USA). The membranes were blocked with TBST buffer (TBS plus 0.1% Tween-20) containing 5% w/v non-fat milk, and hybridized with IRX1 antibody (Code NO. Ab180860, Abcam, USA) at 1:1500 dilution, followed by incubation with specific HRP-conjugated secondary antibody (1:2500, Bioworld, St. Louis, USA). Proteins were visualized using ECL detection system (Amersham Bio-sciences, Uppsala, Sweden).

The 54 glioma only specimens were subject to IHC analysis, by DAB detection kit (Streptavidin-Biotin; ZSGB-BIO, Beijing, China), to assess the expression and location of IRX1. Briefly, following a peroxidase block with 3% H₂O₂/methanol for 30 min, specimens were blocked with 5% normal goat serum. Then the slides were incubated overnight, at 4°C, with mouse polyclonal antihuman IRX1 primary antibody (Code NO. BS2291, Bioworld, USA), at 1:50 dilution. Then, the specimens were briefly washed for 3 times with PBS and incubated at room temperature with the anti-goat secondary antibody for 1 h, followed by incubation with HRP/Fab polymer conjugate for 30 min, at 37°C. The signal visualization was performed by treating with DAB chromogen for 2 to 3 min. After water wash, specimens were counterstained with Meyer's hematoxylin (ZSGB-BIO).

Tissue microarray assay (NO. BS17017b, alenabio) was employed to further validate the IHC results. A total of 62 brain tissue samples were included in the assessment and were classified as; 24 low-grade gliomas [3 pilocytic astrocytomas (WHO I) and 21 diffuse astrocytomas (WHO II)], 33 high-grade gliomas [9 anaplasia astrocytomas (WHO III) and 24 primary glioblastomas (WHO IV)] and 5 adjacent normal brain tissue. All the samples were provided and analyzed, through IHC, by Alenabio biotechnology LTD, Xi'an, China. The immunostaining was performed using mouse polyclonal anti-human IRX1 primary antibody (Bioworld, USA), at 1:50 dilution, while the same was replaced by non-immune IgG antibody for negative control samples. All specimens were counterstained with Meyer's hematoxylin.

The qualitative and semi-quantitative evaluation of the IHC and tissue microarray samples were achieved as follows: five random zones of the stained glioma samples were observed under the light microscope (high power field) and assessed the location of the stain, intensity of staining and the percent of positively stained cells. Based on the intensity and the percent of positively stained cells the samples were scored as follows: no staining as negative (−), <25% positive cells as weakly positive (+), 25%-49% of positive cells as moderately positive (++), and >50% of positive cells as strong positive (+++/++++). The results were evaluated according to WHO classifications by two experienced (>10 years) pathologists, with differences resolved through careful discussion.

The clinico-pathological features of the participating patients were collected from their respective clinical records and the histopathological classification of their glioma condition (WHO classification) was re-evaluated using the study specific H&E staining. The data were correlated to the IRX1 expression and are listed in Table I. The pathological variables affecting the prognosis of the glioma condition were also assessed, through univariate and multivariate analysis.

Further, the survival time of all the 54 patients, from whom glioma only specimens were obtained, was assessed by retrospectively following up for a median period of 60 months. The overall survival time, defined as the time from the date of the initial surgery to death, was calculated based on the Kaplan-Meier method. Patients who died of reasons not related to glioma were excluded from the analysis.

SPSS 17.0 software (SPSS Inc., Chicago, IL, USA) was applied for all statistical analyses, and a P-value<0.05 was considered as significant.

As per the RT-PCR results, all the glioma cell lines (U373, LN229 and T98G cells), except U87 cells, expressed IRX1, while the same was not detected in normal brain tissue (Fig. 1). This suggests of a probable positive correlation of IRX1 with glioma.

Western blot assay showed an increased expression of IRX1 protein in glioma tissue, compared to adjacent normal brain tissue, with the difference reaching statistical significance (P<0.05) (Fig. 2). The data was consistent with our finding in cell lines, which further confirmed a correlation of IRX1 with glioma.

Histologically, among 54 cases of glioma, 25 cases were classified as low-grade gliomas [6 pilocytic astrocytomas (WHO I) and 19 diffuse astrocytomas (WHO II)], and 29 cases as high-grade gliomas [12 anaplasia astrocytomas (WHO III) and 17 primary glioblastomas (WHO IV)]. The IHC analysis of glioma samples revealed that the IRX1 protein was expressed in all grades of glioma tissues. Interestingly the localization of IRX1 shifted from cytoplasm in WHO grade I glioma tissue to nucleus in the higher grade gliomas (Fig. 3A). Further, a positive correlation between IRX1 expression and the glioma grade was observed.

Our IHC results were reconfirmed through tissue microarray (NO.IHC 160917), which, in addition to confirming the expression of IRX1 in glioma tissues, also revealed its presence in the adjacent normal brain tissue. A clear shift in the location of IRX1 from cytoplasm in adjacent normal brain tissue and WHO grade I glioma tissue to nucleus in the higher grade gliomas was observed (Table IIA and Fig. 4), where it took the form of rough trachychromatic claybank granules. Interestingly, one sample in the WHO grade II glioma specimen displayed co-localization of IRX1 in both cytoplasm and nucleus. A few glioma samples with no IRX1 expression was also seen, the number of which sequentially increased with the increase in glioma grade. Among the assessed glioma tissue samples, WHO grade I glioma tissues were restricted to weakly positive IRX1 expression while few WHO grade II and III samples managed to increase IRX1 expression to moderate positivity. Compared to the other glioma grades, significant number of WHO grade IV glioma tissues displayed moderately positive IRX1 expression, with one sample even managing to display strong positivity (Table IIB). The localization difference in tissue microarray analysis was more obvious than that in IHC, and the testing results of normal brain tissues were added in tissue microarray, which is a supplement and further verification of IHC. These results reconfirmed a possible correlation of IRX1 expression with glioma grades.

Univariate analysis showed that large tumor diameter (P=0.04), intra-tumor necrosis (P=0.02) and high IRX1 expression (P=0.008) were significantly associated with poor prognosis. Further multivariate analysis, based on Cox proportional hazards model, identified intra-tumor necrosis (P=0.04) and IRX1 expression (P=0.01) as two independent prognostic variables. Statistical values of the expression of IRX1 and other clinical parameters derived from Cox stepwise proportional hazards model are indicated in Table III.

In total, all the 54 patients had a complete follow-up record. The Kaplan-Meier analysis demonstrated a significant association of IRX1 expression with the overall survival of glioma patients (Fig. 3B).