PSMB9 Codon 60 Polymorphisms Have No Impact on the Activity of the Immunoproteasome Catalytic Subunit B1i Expressed in Multiple Types of Solid Cancer

The proteasome is a key regulator of cellular protein homeostasis and is a clinically validated anticancer target. The immunoproteasome, a subtype of proteasome expressed mainly in hematopoietic cells, was initially recognized for its role in antigen presentation during the immune response. Recently, the immunoproteasome has been implicated in several disease conditions including cancer and autoimmune disorders, but many of the factors contributing to these pathological processes remain unknown. In particular, the codon 60 polymorphism of the PSMB9 gene encoding the β1i immunoproteasome catalytic subunit has been investigated in the context of a variety of diseases. Despite this, previous studies have so far reported inconsistent findings regarding the impact of this polymorphism on proteasome activity. Thus, we set out to investigate the impact of the PSMB9 codon 60 polymorphism on the expression and activity of the β1i immunoproteasome subunit in a panel of human cancer cell lines. The β1i-selective fluorogenic substrate Acetyl-Pro-Ala-Leu-7-amino-4-methylcoumarin was used to specifically measure β1i catalytic activity. Our results indicate that the codon 60 Arg/His polymorphism does not significantly alter the expression and activity of β1i among the cell lines tested. Additionally, we also examined the expression of β1i in clinical samples from colon and pancreatic cancer patients. Our immunohistochemical analyses showed that ∼70% of clinical colon cancer samples and ∼53% of pancreatic cancer samples have detectable β1i expression. Taken together, our results indicate that the β1i subunit of the immunoproteasome is frequently expressed in colon and pancreatic cancers but that the codon 60 genetic variants of β1i display similar catalytic activities and are unlikely to contribute to the significant inter-cell-line and inter-individual variabilities in the immunoproteasome activity.


Introduction
The proteasome is responsible for the degradation of targeted proteins and is a key player in the maintenance of cellular protein homeostasis and the regulation of cellular processes that are essential in cancer development and progression [1,2]. The immunoproteasome, an alternative form of the proteasome, is found in cells of hematopoietic origin, but its expression can also be induced under inflammatory and stress conditions in other cell types [3]. The immunoproteasome is formed when the three types of catalytic subunit found in the constitutive proteasome, b1 (Y, PSMB6), b2 (Z, PSMB7) and b5 (X, PSMB5), are replaced by three homologous immuno-subunits: b1i (LMP2, PSMB9), b2i (MECL-1, PSMB10) and b5i (LMP7, PSMB8). Compared to the constitutive proteasome, the immunoproteasome is found to have slightly altered proteolytic specificities which are capable of producing peptides more suitable for binding to the major histocompatibility complex I molecules thus facilitating antigen presentation [4]. However, recent studies indicate that the immunoproteasome may have important functions beyond the adaptive immune response. For instance, the immunoproteasome is found to be expressed in non-inflammed, immune-previleged tissues (e.g. retina and brain [5,6]) and in the context of several disease states (e.g. cancer, neurodegenerative diseases, autoimmune diseases, [3,7] and references therein). Despite these observations, the biological significance of the immunoproteasome in such disease states is not fully understood.
The b1i subunit of the immunoproteasome is encoded by the PSMB9 gene located on chromosome 6. This gene harbors a commonly occurring genetic R/H polymorphism at codon 60 (p.60R.H; c.179G.A; rs17587) with H allele frequencies of 1.1% to 34%, varying across ethnic groups ( [8] and references therein). Several investigations have reported potential associations between codon 60 PSMB9 polymorphism status and increased susceptibility to various diseases such as insulin-dependent diabetes mellitus, rheumatoid arthritis and multiple sclerosis [8][9][10][11][12][13][14][15][16]. However, it is difficult to decipher from previous studies whether the observed associations are directly related to the altered activity of the b1i subunit as a result of the R/H amino acid variation. This is in part due to the inter-dependent nature of proteasome subunits and the lack of appropriate molecular probes. A previous investigation by Mishto et al. [17] examined more closely at the impact of this polymorphism on proteasome activity in the aged brain using the fluorogenic peptide substrate N-Succinyl-Leu-Leu-Val-Tyr-AMC (Suc-LLVY-AMC). The results from this study suggested that the H allele results in a decreased proteasome activity in the aged brain [17]. However, since Suc-LLVY-AMC is conventionally used to measure the overall chymotrypsin-like (CT-L) proteolytic activity of the proteasome and thus can be hydrolyzed by multiple subunits of both the immunoproteasome and the constitutive proteasome, this decrease in the hydrolysis of Suc-LLVY-AMC may not necessarily indicate changes in b1i function. Furthermore, a subsequent study by the same group using recombinant peptides mimicking endogenous substrates indicated no differences in the substrate hydrolysis profiles between the codon 60 genotypes [18].
Much of the discrepancy regarding the functional impact of the PSMB9 codon 60 polymorphism arises from the lack of a tool to specifically observe the function of the b1i subunit. In this regard, Lin et al. [19] and Blackburn et al. [20] recently reported on the development and application of the fluorogenic substrate Acetyl-Pro-Ala-Leu-7-amino-4-methylcoumarin (Ac-PAL-AMC) which is hydrolyzed selectively by b1i. This novel tool has made it possible to assess the direct functional impact of PSMB9 codon 60 polymorphism on the b1i subunit. In our current study, we investigated the expression of b1i in clinical colon and pancreatic cancer tissues as well as in established human cancer cell lines. Using the b1i-selective fluorogenic substrate Ac-PAL-AMC, we also assessed the catalytic activity of b1i in multiple cancer cell lines carrying different genotypes at codon 60. Our results indicated that b1i is frequently expressed in colon and pancreatic cancers, but the codon 60 PSMB9 polymorphism has no significant impact on the catalytic activity of b1i expressed in multiple types of cancer cell lines.

Immunohistochemical Analysis
Tissue microarrays containing de-identified, archival cases of human colon and pancreatic cancer tissue specimens were obtained from US Biomax. The use of de-identified tissue array samples from a commercial source for the current study was deemed to be exempt from the Human Subject Regulation by the Institutional Review Board. A streptavidin-biotin-immunoperoxidase assay was performed after the antigen retrieval procedure (citrate buffer, pH 6) using a monoclonal antibody against b1i (1:100 dilution, Enzo Life Sciences) according to the previously reported protocol [23]. Immune reaction was visualized using 3,39-diaminobenzidine (DAB), and nuclei were counterstained with hematoxylin. The specificity of immunoreactive signals was verified by omitting either the primary or the secondary antibody. The immunostained tissue microarray sections were analyzed by an experienced pathologist (Dr. Eun Y. Lee). The intensity of immunostaining was assigned on a scale of 0 to 2 or 3.

Determination of PSMB9 Polymorphic Status at Codon 60
The PSMB9 genotypes at codon 60 were initially determined using standard PCR methods and DNA enzymatic digestions as previously reported [24]. The amplicon covering the entire open reading frame was subsequently cloned and analyzed by direct sequencing to verify that the cell lines do not contain additional genetic variations in the PSMB9 gene.

Immunoblotting Analysis
An equivalent amount of tissue or cell extracts were resolved in polyacrylamide gels and transferred to PVDF membranes. Membranes were blocked in 5% skim milk and probed with anti-b1i (1:1000, Abcam) and anti-b-actin (1:1000, Novus) antibodies. After washing, the blots were incubated with horseradish peroxidase-conjugated secondary antibodies. Bound antibodies were detected using an enhanced chemiluminescence substrate (Pierce Biotechnology). In order to compare the relative expression levels of b1i among multiple cell lines, serially diluted H23 cell extracts were used as calibration standards and band intensities were densitometrically quantified by using the Quantity One software (Bio-Rad).

Proteasome Activity Assays
The catalytic activity of b1i was determined by monitoring the hydrolysis rate of fluorescent 7-amino-4-methylcoumarine (AMC) from Ac-PAL-AMC [20]. Briefly, purified proteasomal preparations or cell extracts containing equivalent total protein amount (10 mg) were added to 96 well plates and adjusted to a final volume of 50 ml using assay buffer (20 mM Tris/Cl, 0.5 mM EDTA, pH 8.0). The reaction was initiated by adding Ac-PAL-AMC (100 mM) and the fluorescence of liberated AMC was monitored for 90 min using a SpectraMax M5 plate reader (Molecular Devices, excitation 360 nm and emission 460 nm). In order to verify that substrate hydrolysis is mediated by b1i, additional sets of cell extracts or purified proteasomal preparations were pretreated with proteasome inhibitors UK101 or epoxomicin for 1 h, after which fluorescent signals were monitored. In separate experiments, the hydrolysis of Suc-LLVY-AMC (100 mM) was monitored to assess the overall CT-L proteolytic activity.

Statistical Analysis
Values from data were expressed as means with standard deviations. A comparison between the groups was performed using the Student's t-test and p,0.05 was considered significant.

Frequent Expression of b1i in Colon and Pancreatic Cancer Tissues
In assessing the expression of b1i in colon cancer, we first utilized four pairs of colon cancer and nonmalignant adjacent colonic tissues from matching donors. Our immunoblotting analyses indicated that b1i levels were highly elevated in all four clinical colon cancer tissues compared to the paired nonmalignant colonic tissues ( Figure 1A). To further evaluate the frequency of b1i expression in colon cancer, we performed immunohistochem- ical analyses on a tumor array containing 153 evaluable colon cancer specimens of all clinical stages. The intensity of b1i staining was evaluated on a scale of 0 to 3 and our results indicated that the majority (approximately 70%, n = 107 out of 153 total specimens) of colon cancer tissues were positive (the staining intensity $1) for b1i staining ( Figure 1B). Similar analyses were performed using a tumor array containing 43 evaluable pancreatic cancer specimens of all clinical stages. Due to the limited sample size, the intensity of b1i staining was evaluated on a scale of 0 to 2 and our results indicated that approximately 53% (n = 23 out of 43 total specimens) of pancreatic cancer tissues had positive (the staining intensity $1) b1i staining ( Figure 1C). For both tissue arrays, we assessed whether there is any association between the b1i expression and available clinicopathologic factors (e.g. gender and tumor grade). However, the results did not reveal any apparent associations.

Catalytic Activity of b1i in a Panel of Human Cancer Cell Lines Carrying Different Codon 60 Polymorphic Variants
Previously, Blackburn et al. [20] reported that Ac-PAL-AMC ( Figure 2A) can be used as a selective probe for measuring the catalytic activity of b1i. Prior to the use of Ac-PAL-AMC in our study, we further verified that the hydrolysis of Ac-PAL-AMC was selectively mediated by b1i using purified preparations of immunoproteasome and constitutive proteasome. Indeed, Ac-PAL-AMC was readily hydrolyzed by the immunoproteasome, but not by the constitutive proteasome ( Figure 2B). Pretreatment with UK101, a b1i-selective proteasome inhibitor [21], completely inhibited Ac-PAL-AMC hydrolysis, further validating Ac-PAL-AMC as a b1i-selecitve probe ( Figure 2B). Similarly, approximately 90% of hydrolysis of Ac-PAL-AMC was inhibited in Panc-1 cell extracts by UK-101 pre-treatment ( Figure 2C). In contrast, the hydrolysis of Suc-LLVY-AMC was only partially blocked by UK101 pretreatment in Panc-1 extracts, suggesting that Suc-LLVY-AMC is hydrolyzed by proteasome subunits other than b1i (i.e. b5 or b5i) ( Figure 2C).  Following the validation of Ac-PAL-AMC as a b1i-selective probe, we then assessed the b1i activity in a panel of human cancer cell lines. These cell lines were screened for their PSMB9 genotypes at codon 60 (n = 6 for those carrying HH or HR genotype; n = 11 for those carrying RR genotype) and verified not to have any additional variations in the PSMB9 gene by direct sequencing (Table S1). Our results indicated that the catalytic activities of b1i assessed by the hydrolysis rate of Ac-PAL-AMC substantially vary among these cell lines, however these differences do not seem to be associated with the codon 60 polymorphic status (Figure 3).

b1i Expression in Human Cancer Cell Lines Carrying Different Codon 60 Polymorphic Variants
As a next step, we examined whether the b1i expression levels are influenced by the codon 60 polymorphic status in the tested cancer cell lines. In order to quantify b1i protein levels, which display substantial variability across different cell lines, we used serially diluted H23 cell extracts as calibration standards (Figure 4).
A comparison of cell lines with the H-allele to those that lack the H-allele showed no statistically significant differences in b1i expression ( Figure 5A). Instead, we found a highly significant correlation between the catalytic activity b1i and its expression level ( Figure 5B, r 2 = 0.86, p,0.0001, individual values are included in Table S1). These findings suggested that the differing expression levels of b1i likely account for the majority of variability observed in the b1i activity across different cell lines.

Discussion
In our current study, we investigated the impact of the PSMB9 codon 60 polymorphisms on the b1i subunit of the immunoproteasome expressed in multiple cancer cell lines using a recently reported b1i activity probe, Ac-PAL-AMC. Our results indicated that the codon 60 polymorphism has no major impact on the expression levels or catalytic activities of b1i in the panel of human cancer cell lines tested. These results are different from previous findings which indicated that brain tissue from aged individuals carrying the codon 60 H allele had decreased proteasome activity compared to tissue from those who did not carry the codon 60 H allele [17]. A possible reason for these apparent discrepancies may be related to the different substrates used to assess proteolytic activity. The fluorogenic substrate Suc-LLVY-AMC used by Mishto et al. [17] is commonly used to measure the overall chymotrypsin-like activity of the proteasome. However, since this substrate can be hydrolyzed by both the immunoproteasome (b1i and b5i) and the constitutive proteasome (b5) [25], the results obtained using Suc-LLVY-AMC cannot tease out the direct contribution of the b1i genetic polymorphism. Our current results were obtained using the recently reported substrate Ac-PAL-AMC that is selectively cleaved by the b1i subunit ( Figure 2) [20]. Our findings are consistent with the previous report that the PSMB9 codon 60 polymorphism had no impact on the degradation profiles of a 28-mer peptide (Kloe 258) and a recombinant form of IkBa, a key regulator of classical NF-kB pathway and well-known proteasome substrate [18]. Although we cannot completely rule out the possibility that the impact of the codon 60 polymorphism may vary depending on the substrate and disease types, the results from the current study and others suggest that the codon 60 polymorphism will not likely contribute to inter-individual variability in b1i catalytic activity levels.
With the remarkable clinical successes of bortezomib and carfilzomib in treating multiple myeloma and other hematological malignancies, the proteasome is now recognized as an important chemotherapeutic target. However, undesirable toxicities limit the broad use of the currently available proteasome-targeting drugs. To overcome these limitations, the immunoproteasome, an alternative form of the constitutive proteasome, has been explored as an alternative therapeutic target. Such efforts have yielded promising immunoproteasome-targeting compounds with anticancer efficacy and improved toxicity profiles [7,23,26,27]. In exploring immunoproteasome-targeting approaches for cancer therapy, it is important to consider the immunoproteasome expression across different cancer types. While the immunoproteasome is known to be upregulated in hematological malignancies, the expression of the immunoproteasome in solid cancer has not been thoroughly examined. In the current study, we report that the b1i subunit is frequently expressed in clinical tissue specimens from colon and pancreatic cancers as well as a panel of human cancer cell lines derived from colon, lung, prostate and breast tissues. Although similar findings have been reported by our group and others [21,23,28,29], our current study provides more robust information regarding the expression of b1i in the majority of clinical colon and pancreatic cancer tissues tested. It should be noted that the downregulation of immunoproteasome subunits has also been reported in some types of cancer [30][31][32][33][34] and it is possible that the expression status of the immunoproteasome may be dependent on disease types and their pathogenic mechanisms. On the other hand, it should be noted that our results on the frequencies of b1i polymorphisms of cancer cell lines derived from different organs are not necessarily reflective of those among these cancer types. This is in part due to the small sample size and the experimental design of our current study. In particular, our experimental design involved the exclusion of several cell lines harboring additional polymorphisms in the genes encoding b1i and/or other immunoproteasome subunits such as b5i in order to minimize potential compounding factors. Our results suggest that the codon 60 polymorphisms are not likely to be responsible for the observed variability in the b1i expression/activity among the tested cell lines. In further validating our findings, it may be important to employ a larger sample size of cancer cell lines or clinical samples derived from the same organs, perhaps comparison of samples with comparable b1i expression levels. Additionally, genetic association studies examining the b1i polymorphism at codon 60 in the context of cancer development and progression are also warranted.
In conclusion, our results demonstrate that the b1i subunit of the immunoproteasome is frequently expressed in colon and pancreatic cancers and possibly in other types of solid cancers. In addition, the genetic polymorphism at codon 60 appears to have no major impact on the expression and catalytic activity of b1i in cancer cells. Taken together, these findings may provide useful insights for the development of immunoproteasome-targeting anticancer agents. Additionally, these results exclude codon 60 polymorphic status as a potential factor contributing to variable sensitivity to immunoproteasome inhibitors targeting b1i.