XBP1s levels are implicated in the biology and outcome of myeloma mediating different clinical outcomes to thalidomide-based treatments

X box binding protein 1 (XBP1) is a critical molecule in coordinating normal plasma cell (PC) differentiation and survival due to its role in regulating the unfolded protein response (UPR), correct functioning of which is required for the production of immunoglobulin, the major function of mature PCs.^1-3  Unfolded immunoglobulin is processed within the endoplasmic reticulum where 3 signaling pathways mediated by IRE1α, PERK, and ATF6 are responsible for activating the UPR response.^1,4-8  A build-up of misfolded immunoglobulin results in the oligomerization of IRE1α, inducing autophosphorylation and activation of its kinase domain. This results in an increased IRE1α endoribonuclease activity, which processes a 26 base pair intron from unspliced XBP1 (XBP1u) to form spliced XBP1 (XBP1s), the active transcription factor. XBP1s binds to the promoters of several genes to activate a prosurvival signal mediated by interleukin-6 (IL-6).^8-10  In addition, XBP1s up-regulates other genes involved in the UPR, allowing the mature PC to deal with its immunoglobulin requirements. It would not be a surprise, therefore, if such a key molecule in PC maturation and function were important in myeloma biology and clinical outcome.

Several lines of evidence are consistent with the importance of XBP1 in myeloma pathogenesis.^6,11-13  In XBP1-deficient mice, B cells proliferate and form germinal centers but do not differentiate to immunoglobulin-producing PCs.¹¹  Conversely, transgenic mice expressing high levels of XBP1s develop features characteristic of myeloma.¹⁴  In vitro studies of cell lines show high total XBP1 levels, which can be induced further by IL-6, a growth factor for myeloma cells.¹⁵  The abundant expression of XBP1 in myeloma suggests a role for it maintaining the malignant clone and raises the possibility of it being a prognostic marker and a potential therapeutic target. In particular increased XBP1s levels would be expected to promote PC growth and survival and constitute a poor prognostic factor. In this study we have characterized the IRE1-XBP1 axis in a large series of patients with myeloma. We demonstrate that XBP1s is an independent prognostic variable and show that patients with high XBP1s levels have a poor clinical outcome.

We analyzed 253 patients with newly diagnosed disease from the MRC myeloma IX study. The study compared a thalidomide combination (cyclophosphamide-thalidomide-dexamethasone [CTD]) with standard chemotherapy induction (cyclophosphamide-vincristine-adriomycin-dexamethasone [CVAD] or melpharone-prednisone [MP]). Baseline characteristics are summarized in Table 1 and were similar to that of the whole trial.¹⁶  The median follow-up was 53 months with a median progression-free survival (PFS) of 18 months (range, 0 to ≥ 60 months) and a median overall survival (OS) of 44 months (range, 0 to ≥ 60 months). The trial was approved by the United Kingdom Research Ethics committee (MREC 02/8/95, ISRCTN68454111).

After informed consent was obtained in accordance with the Declaration of Helsinki, RNA was obtained from CD138⁺-selected bone marrow (BM) PCs obtained at diagnosis.¹⁷  Baseline levels of expression of IRE1α and XBP1 were determined by Affymetrix U133Plus2.0 expression levels.¹⁷  A novel Lux–polymerase chain reaction (PCR) method was used to distinguish active XBP1s and inactive XBP1u (Invitrogen). Primers sequences were XBP1-sF(FAM-tag), CACCTCTGAGTCCGCAGCAGG[FAM]G; XBP1-uR(JOE-tag), CAGAGGCTCAGACTACGTCACCTC [JOE]G; XBP1commonR (unlabeled), CCAGAATGCCCAACAGGATA. Thermal cycling conditions were 10 minutes at 95°C, 40 cycles at 95°C for 15 seconds followed by 1 minute at 60°C. Samples were analyzed on a 7500 Fast Real-Time PCR machine (Applied Biosystems).

Correlation analysis between expression levels and continuous variables were determined with the use of Pearson/Spearman correlation. χ² and Fisher exact tests were used to compare nominal variables and the Mann-Whitney U test for continuous variables. Survival curves were plotted according to the Kaplan-Meier method and compared by the log-rank test.¹⁸  Prognostic factors for OS and PFS were determined by the Cox proportional hazard model. The significance of interaction effects were calculated from the difference in the log-likelihoods derived from the Cox models.

All patients expressed high XBP1 mRNA levels, consistent with a central role for XBP1 in PC differentiation, survival, and immunoglobulin secretion. To determine the clinical significance of XBP1s/u ratios, gene expression levels were correlated with baseline clinical characteristics and outcome. Patients were assigned to quartiles, based on their XBPs/u ratio, and these groups were used to analyze survival differences. After adjustment for other independent risk factors [age, β2 microglobulin (β2M), t(4;14)], patients with a low XBP1s/u ratio (≤ 1.33) have a longer OS compared with those with a higher ratio (P = .03, median, 56 months vs 40 months; HR = 1.75; 95% CI = 1.07-2.85; Figure 1A). In addition, XBP1s/u was shown to be independent of the other known prognostic factors: β2M, International Staging System stage, t(4;14), cytogenetic abnormalities, immunoglobulin type, age, and sex. There was no effect of XBP1s/u ratios on response to treatment or PFS.

The clinical effect of genetic markers often differs, depending on the presence of other prognostic markers. Because β2M is one of the most stable and widely used prognostic factors, we looked at its effect by dividing patients into 4 groups, depending on their XBP1s/u ratio and β2M level. Multivariate analysis showed a significant effect of XBP1s/u on OS in the high β2M group (HR = 1.99; P < .05; 33 months vs 59 months; Figure 1B). In the low β2M group this effect was not significant, although it was in the same direction (HR = 1.5; P = .09).

To devise a clinically useful predictor of survival, we took the most important independent prognostic factors [XBP1s/u, β2M, t(4;14)] and divided patients into 4 groups, depending on the number of prognostic factors present. When adjusted by age, results showed a significant difference among the groups for both OS and PFS (P < .001). (For OS: no adverse factor present = 63 months vs 1 factor = 60 months vs 2 factors = 40 months vs 3 factors = 13 months; for PFS: 36 months vs 19 months vs 16 months vs 11 months; Figure 1C).

To determine whether XBP1s/u could be used as a predictive factor for conventional versus thalidomide-based treatment, patients who received thalidomide were compared with patients who received conventional therapy. A highly significant treatment interaction for OS was seen (chi-square test = 16; P < .001; HR = 6.4; 95% CI = 2.3-17.8; Figure 1D), with thalidomide activity being of increased effectiveness in patients with low XBP1s/u. In patients with a low ratio, patients who received thalidomide had a significantly longer OS than patients with a similar ratio who were treated with conventional therapy (P = .001; HR = 7.5; 95% CI = 2.4-23.5; OS, 60% at 60 months vs 44 months). No such effect was seen in the high XBP1s/u group. A similar effect on PFS was seen (treatment interaction χ² test= 6.9; P = .01; HR = 1.6; 95% CI = 0.9-2.6). The XBP1s/u ratio therefore defines a significant-sized group of patients who derive excellent therapeutic benefit from exposure to thalidomide. It is interesting that this effect carries over to survival after relapse because patients with a low XBP1s/u, having already had a better initial PFS, also have a better outcome at relapse (χ² test = 10.3; P = .001; HR = 5.6; 95% CI = 2.0-15.6; Figure 1E).

Thalidomide is known to modulate myeloma growth and survival cytokine levels within the BM microenvironment, including IL-6 and tumor necrosis factor α (TNFα).^19,20  Because XBP1 induces IL-6,¹⁵  it can be postulated that cells with low XBP1s/u would depend more on the BM to release paracrine IL-6 for their growth and survival. Treatment with thalidomide, which inhibits the paracrine cytokine network, would therefore decrease IL-6, leading to reduced growth and increased apoptosis and, thus, an improved patient survival. Patients with high XBP1s/u are less dependent on BM cytokines; hence, the clinical effect of modulating the BM microenvironment is less. These findings are also consistent with recent data showing the importance of upstream regulators of XBP1 in mediating resistance to immunomodulatory derivatives of thalidomide therapy.²¹ 

This study provides further data to suggest a role for high XBP1s in myeloma pathogenesis and to show that patients with high XBP1s/u ratios have a significantly poorer survival. Importantly, the XBP1s/u ratio acts as an independent prognostic marker. Interestingly, we show a highly significant treatment interaction effect with thalidomide exposure; it being highly therapeutically beneficial in patients with low XBP1s/u.

This work was supported by the Kay Kendall Leukaemia Fund, Department of Health, and the Institute of Cancer Research. We also acknowledge National Health Service funding to the National Institute for Health Research Biomedical Research Centre.

Contribution: T.B. performed research, analyzed data, and wrote the paper; P.W., D.G.d.C., E.L.D., B.A.W., K.B., D.C.J., N.J.D., and W.G. performed research and analyzed data; and G.J.M. and F.E.D. designed research and wrote the paper.

Conflict-of-interest disclosure: The authors declare no competing financial interests.

Correspondence: Faith Davies, Brookes Lawley Bldg, Institute of Cancer Research, 15 Cotswold Rd, Sutton, Surrey, SM2 5NG, United Kingdom; e-mail: faith.davies@icr.ac.uk.