Expression of functional lung resistance–related protein predicts poor outcome in adult T-cell leukemia

Adult T-cell leukemia (ATL) is caused by human leukemia virus type I and is divided into 4 clinical subtypes (acute, chronic, lymphoma, and smoldering).1 Chronic and smoldering ATLs have a mild clinical course and do not require treatment with intensive chemotherapy. Acute and lymphoma-type ATLs require intensive chemotherapy, and the median survival period is less than one year.2 

A major obstacle to successful treatment of ATL is thought to be intrinsic drug resistance.2 Several mechanisms for multidrug-resistance have been elucidated. Enhanced drug efflux by membrane transport proteins, such as P-glycoprotein3and multidrug resistance protein (MRP)–1, MRP2, MRP3, MRP4, and MRP5, contribute to drug resistance.4 Another mechanism of drug resistance is the intracellular redistribution of drugs without changing their intracellular accumulation.5,6 Lung resistance–related protein (LRP) is expressed in non–P-glycoprotein–mediated drug-resistant cell lines,7and it was found to be identical to the major vault protein.8 Mammalian vaults are composed of 3 proteins of 100, 193, and 240 kd and a small untranslated RNA. The 100-kd subunit is the major portion vault protein and constitutes more than 70% of vaults. The 193-kd subunit is a novel poly(adenosine 5′-diphosphate [ADP]-ribose) polymerase—vault poly(ADP-ribose) polymerase9—and the 240-kd subunit is identical to the mammalian telomerase-associated component, telomerase associated protein 1 (TEP1).10 Vaults are cytoplasmic organelles,11 a small portion of which is localized in the nuclear membrane and nuclear pore complex.12 Vaults are thought to be involved in both the vesicular and nucleocytoplasmic transport of drugs.6 Using LRP-specific ribozymes, we recently showed that LRP is involved in drug resistance in sodium butyrate-treated human colon carcinoma SW-620 cells.13 

In this study, we assessed LRP expression in ATL and demonstrated that it does the nucleocytoplasmic transporting activity of doxorubicin (DOX) in the ATL cells and that high expression of LRP correlated with poor outcome in ATL.

Between July 1989 and July 1998, we studied 55 ATL patients (Table 1). They consisted of 26 males and 29 females with a median age of 65.0 years (range, 42-86 years). According to previously reported diagnostic criteria,1 36 of these patients had acute, 12 chronic, and 7 lymphoma-type ATL. Fourteen patients had hypercalcemia (corrected calcium level with serum albumin greater than 2.75 mM [5.5 mEq/L]). Performance status (PS) was based on the 5-grade scale of the World Health Organization. We found PS 0, 1, 2, 3, and 4 in 13, 20, 10, 2, and 10 patients, respectively. During this study, we treated acute and lymphoma-type ATL patients with combination chemotherapy regimens, such as a response-oriented multidrug protocol; a cyclophosphamide, DOX, vincristine, and prednisone protocol; or LSG15 protocol.2In this study, these protocols did not result in significant survival differences. Response was evaluated according to previously reported criteria.2 None of the chronic ATL patients required treatment with intensive chemotherapy. After informed consent, 50 samples (from patients 1 to 35 and 37 to 55) were obtained at the time of diagnosis and 5 samples (from patients 32 to 36) at relapse. Samples 1 to 36 were from acute, 37 to 48 from chronic, and 49 to 55 from lymphoma-type ATL.

Peripheral blood mononuclear cells separated by Ficoll-Conray density gradient centrifugation were collected from 30 patients with acute and 12 patients with chronic ATL. Lymph node samples were taken from 6 patients with acute and 7 with lymphoma-type ATL. All samples contained more than 80% abnormal lymphocytes. The samples were stored at −110°C.

Human epidermoid KB carcinoma cells and human fibrosarcoma HT1080 cells were grown as monolayers in Eagle minimum essential medium (Nissui Seiyaku Tokyo) containing 10% newborn calf serum (Sera-Lab, CrawleyDown, United Kingdom), 1 mg/mL bactopeptone (Difco, Detroit, MI), 0.292 mg/mL glutamine, and 188 μM (100 U/mL) penicillin. KB cells were subcloned twice; a single recloned line, KB-3-1,14 was used as a drug-sensitive parental cell line. KUT-215 is a human lymphotropic virus type 1–infected cell line established from peripheral blood mononuclear cells of an ATL patient. The [¹⁴C]DOX (2.11 GBq/mmol [57.0 mCi/mmol]) was from Amersham International (Buckinghamshire, United Kingdom). All other chemicals were obtained from Sigma (St Louis, MO).

To prepare RNA probes for Northern and slot blotting, the 914–base pair (bp) fragment of human LRP complementary DNA corresponding to LRP nucleotides 1881 through 2794 and a 982-bp fragment (nucleotides 6 through 987) of human glyceraldehyde 3-phosphate dehydrogenase obtained by reverse-transcriptase polymerase chain reaction were inserted into pBluescript II SK⁺ and pT7T3 18U multifunctional Phagemid (Pharmacia, Uppsala, Sweden), respectively. After the vectors were linearized, single-strand RNA probes were prepared by means of a digoxigenin (DIG) RNA labeling kit (Boehringer Mannheim, Mannheim, Germany).

Total cellular RNA was extracted in a single step by means of Trizole Reagent (Life Technology, Rockville, MD). The integrity of isolated RNA was monitored by ethidium bromide staining after gel electrophoresis. For slot blotting, 1 μg and 5 μg RNA were applied to a nylon membrane (Hybond N⁺) (Amersham) under vacuum; the membrane was dried at room temperature and then fixed with UV irradiation. The membrane was prehybridized for 1 hour at 68°C in 50% formamide, 5 × SSC (standard saline citrate) (1 × SSC = 0.15 M NaCl, 15 mM sodium citrate, pH 7.0), and 1% sodium dodecyl sulfate (SDS), containing 0.05 mg/mL yeast transfer RNA (tRNA). The membrane was then hybridized for 16 hours at 68°C in 50% formamide, 1% SDS, and 5 × SSC containing both 0.05 mg/mL tRNA and 100 ng/mL DIG-labeled RNA probe.

After hybridization, the membrane was washed twice for 5 minutes at room temperature with 2 × SSC containing 0.1% SDS and then twice for 60 minutes at 68°C with 0.1 × SSC containing 0.1% SDS. Hybridization was detected by means of alkaline phosphatase–labeled anti-DIG antibody and chemiluminescent substrate according to the manufacturer's instructions.

The density of the LRP messenger RNA (mRNA) signal on the autoradiograms was determined by means of a GS-525 Molecular Imager System (Bio-Rad, Hercules, CA), and the density of the signal for 5 μg total RNA from KB-3-1 cells was arbitrarily assigned a value of 1 U. The data are shown relative to LRP mRNA in KB-3-1 cells.

For Northern blotting, poly(A)⁺ RNA was extracted from total RNA with a Bio Mag mRNA Purification Kit (Perseptive Biosystem, Framingham, MA). Poly(A)⁺ RNA (1 μg) was resolved by electrophoresis on a formaldehyde-agarose denaturing gel. After blotting onto a nylon membrane (Hybond N⁺), hybridization was carried out with the same protocol as used for slot blot analysis.

To prepare total cell lysates, cells were washed with phosphate-buffered saline (PBS) and scraped into PBS containing 1% Igepal CA-630 (Sigma), 0.5% sodium deoxycholate, 0.1% SDS, 1% aprotinin (Sigma), and 1 mM P-amidinophenyl methanesulfonyl fluoride hydrochloride (Wako, Osaka, Japan). The lysates were passed through a 21-gauge needle to shear the DNA, incubated for 30 minutes on ice, and centrifuged at 15 000g for 20 minutes at 4° C. The supernatants were stored at −80°C.

Nuclei were isolated as previously described16 and suspended in solution A (250 mM sucrose, 1 mM dithiothreitol, 80 mM KCl, 15 mM NaCl, 5 mM EDTA, 15 mM piperazine diethanesulfonic acid–NaOH [pH 7.4], 0.5 mM spermidine, 0.2 mM spermine, and 1 mM phenylmethylsulfonyl fluoride).

We used a rabbit anti–human LRP polyclonal antibody13 to detect LRP. Protein concentrations were determined by the procedure of Bradford.17 We mixed 50 μg proteins from total cell lysates and 10 μg proteins from the isolated nuclei with an equal volume of SDS sample buffer consisting of 125 mM Tris-HCl (pH 6.8), 4% SDS, 20% glycerol, and 0.005% bromophenol blue. The proteins were resolved by electrophoresis on SDS 7.5% (wt/vol) polyacrylamide minigels and transferred to polyvinylidene difluoride (PVDF) membranes, Immobilon-P (Millipore, Bedford, MA), by means of a Transblot SD apparatus (Bio-Rad). Thereafter, the membranes were blocked with 3% skimmed milk in buffer C (0.35 M NaCl, 10 mM Tris-HCl [pH 8.0], and 0.05% Tween 20) for 1 hour at room temperature and then incubated for 4 hours with 5000-fold diluted antibody against LRP in buffer C containing 3% skimmed milk. The membranes were washed 3 times with buffer C and then incubated for 1 hour with 1000-fold diluted anti–rabbit immunoglobulin whole antibody labeled with horseradish peroxidase (Amersham). The membranes were rinsed once for 15 minutes and 4 times for 5 minutes with buffer C and then evenly coated by means of the enhanced chemiluminescence (ECL) Western blotting detection system (Amersham) for 1 minute. The membranes were then immediately exposed to Fuji medical x-ray film Rx-U (Fuji photo film, Tokyo, Japan) in a film cassette at room temperature for various periods.

To examine [¹⁴C]DOX accumulation in isolated nuclei, 1 × 10⁵ nuclei were incubated at 37°C for 2 minutes with 1 μM [¹⁴C]DOX in solution A, in either the absence or the presence of anti-LRP polyclonal antibody (15 μg/mL) and normal rabbit serum (15 μg/mL). The nuclei were then washed 3 times with solution A, and the radioactivity of the nuclei was determined. The accumulation of DOX in the isolated nuclei was also examined by fluorescence microscopy. The isolated nuclei were suspended in solution A and incubated with 1 μM DOX at 37°C for 2 minutes in the presence or absence of anti-LRP polyclonal antibody or normal rabbit serum. The nuclei were then washed with solution A and examined by fluorescence microscopy (Olympus, Tokyo, Japan). To study [¹⁴C]DOX efflux, 1 × 10⁵ nuclei were incubated at 37°C for 10 minutes with 1 μM [¹⁴C]DOX in solution A with and without anti-LRP antibody, washed once with ice-cold solution A, and then incubated at 37°C. After 10 and 20 minutes with and without anti-LRP polyclonal antibody, the nuclei were washed 3 times with solution A, and radioactivity of the nuclei was determined.

Differences between groups were analyzed by one-way analysis of variance or Student t test. A P < .05 was considered significant. We used the Kaplan-Meier method to estimate survival and the log-rank test to compare the groups for differences in survival. Survival durations were measured from time of chemotherapy to time of death. The Cox proportional hazards model was used in the univariate and the multivariate survival analyses. Maximum likelihood parameter estimates and likelihood ratio statistics were obtained. We calculated Wald-type confidence intervals. All P values presented were 2-sided.

The levels of LRP mRNA in 55 ATL samples, the KB cell lines, and KUT-2 cells were examined by slot blotting. The results of typical slot blots are shown in Figure 1A. The density of the signal for 5 μg total RNA from KB-3-1 cells was arbitrarily assigned a value of 1 U. ATL cells from patients 17, 24, 29, 31, 32, 45, and 52 expressed 4.09-, 10.12-, 14.06-, 22.23-, 3.80-, 6.03-, and 2.44-fold higher levels of LRP mRNA, respectively, than KB-3-1 cells. A 3.5-kilobase (kb) LRP mRNA was detected by Northern blots in the samples examined (patients 20 and 52) and in HT1080 cells in which the expression levels were higher than in drug-sensitive KB-3-1 cells (Figure 1B). The expression levels of LRP mRNA in these cases correlated with the levels of LRP protein (Figure 1A).

The expression levels of LRP mRNA in all ATL samples were quantified. The expression levels ranged from 0.49 to 22.23 U (median, 4.77 U). The expression levels were significantly higher in chronic ATL than in lymphoma-type ATL (P = .007) (Figure2A). Significant associations were observed between LRP mRNA expression and white blood cell (WBC) counts (P = .038) or abnormal lymphocyte counts (P = .007) (Figure 2B-C). The expression of LRP mRNA was higher in patients with WBC counts above 30 × 10⁹/L (30 000/μL) or with abnormal lymphocyte counts above 10 × 10⁹/L (10 000/μL) than in the rest. However, the expression of LRP mRNA did not correlate with age, gender, lactate dehydrogenase (LDH), hypercalcemia, blood urea nitrogen (BUN), or PS. There was no significant difference between LRP mRNA expression at diagnosis and at relapse (data not shown). In recurrent patients, the expression of LRP mRNA did not correlate with age, gender, subtype, WBC, abnormal lymphocyte counts, LDH, hypercalcemia, BUN, or PS.

To investigate whether the LRP expressed in ATL cells is functional, we examined the accumulation and efflux of DOX in nuclei isolated from ATL cells (Figure 3). The expression of LRP in isolated nuclei was examined by immunoblot analysis (Figure 3A). The expression levels of LRP mRNA from patients 28, 29, 31, 32, 41, 42, 46, and from KUT-2 cells were 13.27, 14.06, 22.23, 3.80, 6.31, 13.98, 13.30, and 0.73 U, respectively. The expression levels of LRP protein in the isolated nuclei were correlated with those of LRP mRNA in the ATL cells from which the nuclei were isolated.

When the nuclei were incubated with [¹⁴C]DOX alone, the accumulation of [¹⁴C]DOX in the nuclei isolated from ATL cells was lower than that from KUT-2 cells and inversely correlated with the expression levels of LRP. The accumulation of [¹⁴C]DOX in the nuclei isolated from patients 28, 29, 31, and 42, but not from KUT-2 and patient 41, was significantly enhanced by incubation with an anti-LRP antibody (P = .028,P = .043, P = .034, andP = .033, respectively) (Figure 3B). However, the [¹⁴C]DOX accumulation in the nuclei from these ATL cells was not significantly enhanced by incubation with normal rabbit serum (data not shown).

The effluxes of [¹⁴C]DOX from the nuclei isolated from ATL samples (patients 29 and 31) that expressed high levels of LRP mRNA were significantly enhanced compared with nuclei from KUT-2 cells (P = .002 and P < .001, respectively). The polyclonal antibody against LRP partially inhibited the enhanced effluxes from the nuclei isolated from the ATL cells (patients 29 and 31) (P = .019 and P = .038, respectively) (Figure 3C).

The accumulation of DOX in the isolated nuclei was examined by fluorescence microscopy (Figure 3D). When the nuclei were incubated with DOX alone, its accumulations in the nuclei from ATL patients 28, 29, 31, and 32 were lower than from KUT-2 cells. However, DOX accumulation in the nuclei from the ATL samples was enhanced by incubation with anti-LRP antibody, and results were similar in KUT-2 cells. The accumulation of DOX in the nuclei from KUT-2 cells was not increased in the presence of anti-LRP antibody.

The prognostic relevance of LRP expression was studied in 38 patients with acute or lymphoma-type ATL at diagnosis. We divided the patients at diagnosis into 3 groups according to their LRP mRNA expression levels and compared their survival periods (Figure4). Levels were regarded as low when LRP mRNA expression was less than 3 U (13 patients); intermediate when the expression was from 3 to 6 U (12 patients); and high when the expression was above 6 U (13 patients). The expression of LRP mRNA was associated with shorter survival. Median overall survival was 6.0, 6.3, and 2.8 months for patients with low, intermediate, and high LRP mRNA expression, respectively. Patients with low or intermediate expression had a longer survival than patients with high expression (P = .032). High WBC and high abnormal lymphocyte counts were not associated with shorter survival (data not shown). There was no differences in significant clinical features between high and low patients.

Next, we investigated whether LRP expression is an independent prognostic factor. We constructed a Cox proportional hazards model using established prognostic factors19 and LRP expression. In the univariate analysis (Table 2), hazard ratios for death were 3.09 for high BUN (P = .009), 2.20 for high PS (P = .032), 2.69 for high LRP mRNA expression (P = .021), 1.03 for age (not significant [NS]), 1.46 for LDH of 1500 IU/L or greater (NS), and 1.65 for hypercalcemia (NS). In the univariate analysis using LRP mRNA expression as a continuous variable, the expression of LRP mRNA was associated with shorter survival (P = .030). Next, we performed a multivariate analysis that included all parameters significantly associated with outcome in the univariate analysis. In the multivariate analysis (Table 2), hazard ratios for death were 4.30 for high BUN (P = .006) and 4.07 for high LRP mRNA expression (P = .008). Although only the highest category of LRP among the 3 categories showed a significant increase of hazard ratio, a trend test using LRP expression as a continuous variable showed a significant increasing trend of hazard ratio (1.13) with the elevation of LRP mRNA levels (P = .029).

In the present study, we have demonstrated that LRP is expressed in ATL and that high LRP expression correlated with shorter survival in acute and lymphoma-type ATL. Although LRP mRNA expression in ATL was associated with WBC and abnormal lymphocyte counts (Figure 2), WBC and abnormal lymphocyte counts had no impact on outcome of survival in our study (data not shown). Thus, the survival difference we found among low, intermediate, and high LRP patients cannot be explained by a high WBC or high abnormal lymphocyte counts in patients with high LRP. Survival was not significantly longer for lymphoma-type ATL than for acute ATL (data not shown). Although the expression of LRP mRNA in chronic ATL was significantly higher than in lymphoma-type ATL and nuclei isolated from chronic patients (patients 41 and 42) had functionally active LRP in this study (Figures 2A, 3B), chronic ATL did not require treatment with intensive chemotherapy, and it was unclear whether LRP expression correlates with clinical outcome in chronic ATL.

In the univariate and multivariate survival analyses, we demonstrated that high LRP expression and high BUN were poor prognostic factors in ATL. Age, LDH, hypercalcemia, and PS, which had previously been reported as prognostic factors,18 were not poor prognostic factors in this study. These results strongly suggest that LRP expression is involved in drug resistance in acute and lymphoma-type ATL.

Similar results were reported in other hematological malignancies. LRP is expressed in 36% of the patients with de novo acute myeloid leukemia (AML),19 and the complete remission rate is lower and overall survival is significantly shorter in AML patients with LRP expression than in patients without LRP expression in de novo and secondary AML.19,20 In multiple myeloma,21LRP expression is found in 61% of the patients and is a significant prognostic factor.

The mechanisms of drug resistance mediated by LRP have been unclear until now. We demonstrated that the efflux of DOX from nuclei isolated from ATL cells that overexpressed LRP was enhanced and that this efflux was inhibited by anti-LRP antibody. Moreover, the accumulation of DOX in nuclei isolated from ATL cells that expressed LRP was enhanced by incubating with anti-LRP antibody. These results indicate that functionally active LRP is expressed in ATL cells and that the expressed LRP transports DOX from nuclei to cytoplasm. These results are also consistent with our previous report.13 DOX is mainly distributed in the cytoplasm of cells that express LRP but in the nuclei of cells that have no detectable LRP.13 LRP was shown to be involved in resistance to DOX and to play an important role in the transport of DOX from the nuclei to the cytoplasm.13 These results suggest that LRP-mediated drug resistance is caused by the transport of anticancer drugs from nuclei to cytoplasm.

Raaijmakers et al22 reported that high-dose melphalan overcame LRP-mediated resistance in multiple myeloma. We observed that DOX accumulation in the nuclei isolated from LRP-overexpressing ATL cells increased when they were incubated with a high dose of DOX (data not shown). The possibility of overcoming LRP-mediated drug resistance by increasing the dose was also supported by our findings. We also found that a pyridine analog, PAK104P,23 24 enhanced the accumulation of DOX in the nuclei isolated from ATL cells and at least partially inhibited the efflux of DOX from the isolated nuclei (data not shown). Thus, modulation of LRP-mediated resistance appears to be feasible, and this modulation may improve the clinical outcome of ATL in the future.

In summary, LRP expression was shown to be an independent prognostic factor in ATL associated with shorter survival. Our study confirms the hypothesis that poor outcome in the presence of LRP expression is due to the enhanced efflux of anticancer drugs from nuclei to cytoplasm. Further study of LRP function and its inhibition is needed to overcome LRP-mediated drug resistance in ATL.

We thank R. Pirker for critical reading of the manuscript.