Loss of FHIT Expression in Acute Lymphoblastic Leukemia¹

The FHIT gene at chromosome 3p14.2 spans over 1 Mb and includes the FRA3B common fragile region, the t(3;8) (p14.2;q24) renal cell carcinoma-associated translocation, and a human papillomavirus integration site (1, 2). Distributed over the FHIT genomic locus are 10 small exons encoding a 1.1-kb mRNA. The first four exons are untranslated; exon 5 contains the start codon, and exon 9 contains the stop codon. The FRA3B fragile region encompasses over 500 kb surrounding exon 5, the t(3;8) break lies in intron 3, and the human papillomavirus integration site involves intron 4 (1, 3). The FHIT gene is expressed at varying levels in most adult tissues (1, 4) and encodes a 147-amino acid protein with in vitro diadenosine 5′,5‴-p¹,p³-triphosphate hydrolase activity (5). The physiological function of Fhit is still unknown.

Homozygous deletions, a hallmark of tumor suppressor genes, have been observed in the FHIT gene in head and neck, esophageal, gastric, colon, lung, and cervical cancers or cancer cell lines (1, 6, 7, 8, 9, 10, 11, 12), often including exon 5. Absence or alteration of FHIT transcription has been observed in head and neck, esophageal, gastric, pancreatic, lung, breast, kidney, and cervical carcinomas (4, 6, 7, 13, 14, 15, 16, 17, 18) by RT³-PCR or Northern analysis. Siprashvili et al. (19) have shown that transfection of FHIT into tumorigenic cell lines suppresses tumorigenicity in nude mice, indicating that FHIT is a tumor suppresser gene.

Although chromosome 3 has not been reported to be a primary target of cytogenetic aberrations in leukemias, translocations at 3p14 (20) and loss of heterozygosity at 3p (21) have been reported. Aberrant FHIT transcripts lacking various exons between exons 3 and 9 have been found in acute and chronic leukemias in addition to the wild-type transcript, but total loss of FHIT mRNA expression has been detected infrequently (22, 23, 24, 25, 26). Nevertheless, absence of the Fhit protein has been detected in hematological malignancies (26).

To assess the role of FHIT in ALL, we examined PBLs from patients with ALL and ALL-derived cell lines and PBLs from healthy volunteers and lymphoblastoid cell lines for DNA integrity, transcription, and protein expression of the FHIT gene.

Cancer-derived cell lines were obtained from the American Type Culture Collection. All cell lines were maintained in RPMI 1640 containing 10% FCS and 0.1 mg/ml gentamicin. Uncultured leukemic cell samples from 18 patients with ALL, including 6 T-cell and 12 B-cell leukemias, were obtained from The University of Texas M. D. Anderson Cancer Center, and five samples of PBLs were obtained from healthy volunteers.

Total RNA was isolated from cell lines and PBLs using the RNeasy Mini Kit (Qiagen, Valencia, CA) according to the manufacturer’s instructions, and RT was performed from 2–3 μg of total RNA in a 20-μl final volume of 20 mm Tris-HCl (pH 8.4), 50 mm KCl, 2.5 mm MgCl₂, 10 mm DTT, 0.5 mm dNTPs, 0.5 μg of oligo(dT) primers, and 200 units of reverse transcriptase at 42°C for 60 min. The reaction was terminated at 70°C for 10 min.

Nested RT-PCR amplifications were carried out as described previously (1). Additionally, seminested RT-PCRs were performed under the same conditions using primers 5D2 (exon 2) and 3U2 (exon 10) in the first round and primers 5D2 and 3U1 (exon 10) (1) in the second round of amplification. PCR products were analyzed in a 1.5% agarose gel. Amplification with β-actin primers was performed from every sample to confirm the quality of the cDNA.

PCR-amplified bands were excised from the gels, and DNA was extracted using Qiaquick Gel Extraction Kit (Qiagen) according to the manufacturer’s instructions. DNA (25 ng) was directly sequenced using one of the PCR primers for cycle sequencing and analyzed in ABI 377 automated sequencers.

Total RNA from the cell lines was obtained by extraction with Trizol reagent (Life Technologies, Inc., Grand Island, NY) according to the manufacturer’s instructions. Poly(A) mRNA was isolated from 1 mg of total RNA using oligo(dT) cellulose spin columns (5′→3′, Boulder, CO) as described by the manufacturer. Poly(A) mRNA (3 μg) was separated by electrophoresis in 0.8% denaturing agarose gels, and the quality of the mRNA was checked visually under UV illumination. The mRNA was then transferred to nylon membranes in 20× SSC. The membranes were hybridized with a cDNA probe consisting of exons 2–9 of the FHIT gene labeled with [α-³²P]dCTP by random priming. Prehybridization and hybridization were carried out in 50% formamide, 5× SSPE, 10× Denhardt’s solution, 2% SDS, and 0.1 mg/ml single-stranded DNA at 42°C. Hybridized membranes were washed in 2× SSC, 0.1% SDS, and in decreasing concentrations of SSC for 20 min each at 60°C.

High molecular weight DNA from cell lines and PBLs was obtained using standard phenol-chloroform extraction. DNA (6 μg) was digested with restriction enzymes EcoRV or HindIII in a 40-μl reaction mix containing 1× buffer supplied by the manufacturer and 20 units of enzyme. The DNA was separated by electrophoresis in 0.7% agarose gels and blotted to nylon membranes using the Probe Tech 2 Oncor machine according to the manufacturer’s instructions. The membranes were hybridized with the same probe as the Northern blots. Prehybridization and hybridization were carried out at 65°C in 5× Denhardt’s solution, 5× SSPE. 1% SDS, and 0.1 mg/ml single-stranded DNA. Hybridized membranes were washed in 2× SSC, 0.1% SDS, and decreasing concentrations of SSC (down to 0.1× SSC) for 20 min each.

Cells from cell lines and PBLs were washed in PBS and incubated in lysis buffer [250 mm NaCl, 50 mm Tris (pH 7.5), 1 mm EDTA, 1% Triton X-100, 1 mm DTT, 1 mm phenylmethylsulfonyl fluoride, and 10 μg/ml each of leupeptin and pepstatin] for 60–90 min on ice. Lysates were centrifuged for 15 min at 10,000 × g, and the supernatants were used for further investigation. Protein concentration was measured in a BCA protein assay (Pierce, Rockford, IL). Protein (50–100 μg) was separated in 15% SDS-PAGE gels and blotted onto nitrocellulose Hybond ECL (Amersham, Piscataway, NJ). After blocking in 5% nonfat milk, the membrane was incubated with anti-Fhit rabbit serum (Refs. 9 and 19; diluted 1:4,000 in PBS-Tween) for 1–2 h at room temperature. The secondary antibody was antirabbit immunoglobulin labeled with horseradish peroxidase (Amersham). The signal was detected using the ECL system (Amersham) as described by the manufacturer. The quality of the protein was confirmed by incubating the membranes with anti-tubulin immunoglobulin (Neomarkers, Fremont, CA) as primary antibodies instead of anti-Fhit serum.

Cells from cell lines and PBL samples were washed in PBS and fixed in 4% PBS-buffered formalin for 10 min before drying on glass slides. After blocking with goat serum, the slides were incubated with α-Fhit rabbit serum (Refs. 9 and 19; diluted 1:4000 in PBS with 1% BSA) overnight at 37°C. They were then incubated with biotinylated goat α-rabbit immunoglobulin. The immunoglobulin was detected using the Vectastain ABC Reagent (Vector Laboratories, Burlingame, CA) according to the manufacturer’s instructions; the cells were counterstained with Harris’ Hematoxylin.

Uncultured leukemic cell samples from 18 ALL patients, 8 ALL cell lines, and 7 cell lines derived from other hematological malignancies [2 AMLs, 1 CML, 1 immunoblastic B-cell lymphoma, 1 T-cell lymphoma, 1 Burkitt’s lymphoma, and 1 tumorigenic lymphoblastoid cell line (GM1500–6TG-Oub)] were analyzed by immunocytochemistry and Western blotting. As controls, 14 EBV-transformed lymphoblastoid cell lines and 5 PBL samples from healthy volunteers were also investigated.

Surprisingly, none of the lymphocyte samples showed any staining by immunocytochemistry using anti-Fhit rabbit serum, although Fhit expression was detected by Western blotting and RT-PCR (see next paragraph). Other tissues used as controls (the kidney cell line 293 and normal breast and prostate tissue) displayed clear cytoplasmic staining in the epithelial cells. Inflamed tonsil tissue containing nonmalignant lymphocytes also showed no staining in the B- or T-cell compartments, although a weak staining was difficult to judge due to the small amount of cytoplasm in lymphoblastoid cells. The tonsil epithelium, on the other hand, showed clear cytoplasmic staining (data not shown). These results suggested that hematopoietic cells express low levels of Fhit relative to epithelial cells.

Expression of Fhit protein was detected by Western blot in all lymphoblastoid cell lines and normal PBL samples (Table 1), whereas it was found only in one of eight ALL cell lines and one of seven other hematological cancer cell lines. In the Burkitt’s lymphoma cell line Daudi and the Jurkat cell line (T-ALL), a very faint Fhit band could be detected only after a very long exposure of the blot. Of the primary ALLs, only 6 of 18 cases showed robust expression of Fhit protein that was similar in signal strength to the lymphoblastoid cell lines or normal PBL samples, with the same amount of protein being loaded in every case (cases 2, 3, 8, 9, 13, and 16; see Table 2). After a very long exposure of the blots, using the same conditions as described for the cell lines, faint bands could also be observed in the remaining 12 primary ALL cases (Fig. 1). Because the leukemic cell samples are contaminated with normal PBLs, a low level of Fhit expression is probably due to the nonneoplastic cells in these cases. Interestingly, Fhit expression was found only in B-cell leukemias and not in any of the six T-cell leukemias. The significance of this finding was confirmed using Fisher’s exact test. Therefore, loss of Fhit expression seems to be more common in T-cell leukemias than in B-cell leukemias. Fhit expression in normal lymphocyte samples was lower than in the kidney cell line 293 used as a positive control, because 25 μg of protein from the 293 cell line gave a signal for Fhit protein similar to the signal from 50 μg of protein from lymphoblastoid cell lines or PBL samples, thus confirming a low level of Fhit protein expression in lymphocytes.

In summary, expression of Fhit protein could be detected in only 24% of the malignant samples (including primary ALL cases and tumor cell lines) but was detected in 100% of normal samples (lymphoblastoid cell lines and PBL samples; see Table 1). This difference was highly significant using the χ² test (P < 0.001).

RNA samples from seven ALL-derived cell lines, six cell lines derived from other hematological malignancies, and seven normal lymphoblastoid cell lines were examined for the presence of FHIT mRNA by Northern blotting. FHIT mRNA was found in all normal lymphoblastoid cell lines but in only two of the neoplastic cell lines investigated (Table 2 and Fig. 2). After a long exposure of the blot, the ALL cell line Jurkat and the Burkitt’s lymphoma cell line Daudi each showed a weak band of FHIT mRNA, thus confirming the results obtained by Western blotting.

All samples were also investigated for FHIT mRNA expression by nested and seminested PCR amplification. Samples in which normal FHIT transcripts could not be detected by RT-PCR did not show any FHIT mRNA by Northern blotting or Fhit protein by Western blotting, but in some samples, normal-sized PCR products could be found despite the absence of Fhit protein or FHIT mRNA as detected by Northern blotting.

Amplification of exons 3–10 by nested PCR in all lymphoblastoid cell lines and normal PBL samples showed a FHIT transcript of wild-type length. However, three of eight ALL cell lines and the immunoblastic lymphoma cell line JM displayed no FHIT transcript at all, and in the CML cell line K562, only a smear of numerous aberrant bands was found. The other cell lines, including the lymphoblastoid cell lines and three of five PBL samples, showed one to three faint aberrant bands in addition to the normal-sized band. In only one AML cell line (TMP-1) with a t(9:11) chromosome translocation, the normal band was the only FHIT RT-PCR product detected. Repeated amplification of FHIT cDNA from the primary ALL cases did not necessarily reveal the same products in each experiment. A total of 16 of 18 ALL cases displayed the normal-sized transcript, but in 2 cases (cases 10 and 11), various attempts of FHIT amplification from two different cDNA preparations revealed the normal transcript in some of the reactions, but in others, no amplification product was found. β-Actin PCR of these samples revealed no sign of degradation or poor quality of the cDNA. This result can be explained by a very low amount of FHIT cDNA in the samples. Most of the ALL cases showed various faint, shorter, aberrant bands that were not reproducible with regard to occurrence and length in repeated amplifications. Sequencing of aberrant bands from cell lines or primary ALL samples revealed that in most cases, complete exons were missing, most often exons 4–7 or 4–8.

Sugimoto et al. (22) reported aberrant transcripts most often lacking exons 3–6 in their leukemia cases. Because these aberrant transcripts would not be detected by our PCR with primers placed in exon 3, we developed a seminested PCR amplifying exons 2–10. The five cell lines displaying no normal-sized FHIT transcript in the nested PCR did not show it by this approach either. The primary ALL cases 10 and 11 again showed alternatively positive and negative results for normal-sized FHIT cDNA. As in the nested approach, one to three additional aberrant bands were found in all cell lines except TMP-1 (shown in Fig. 3), which displayed only the normal-sized band, and K562, which showed only a smear of aberrant transcripts. By this approach, the aberrant bands were often as strong as the normal-sized bands, and sequencing revealed that the most prominent aberrant bands lacked exons 3–6 or 3–7, mostly accompanied by fainter bands lacking exons 4–7. Most of the primary ALL cases and three of five PBL samples showed aberrant RT-PCR products.

To summarize, the two different RT-PCR approaches revealed a lack of normal-sized FHIT transcripts in 5 of 15 (33%) tumor cell lines but in none of the normal lymphoblastoid cell lines (Table 2), with the difference being significant (P < 0.05). Additionally, aberrant transcripts were found in all cell lines with the exception of one AML-derived cell line (TMP-1).

The sequences of 10 randomly chosen normal-sized RT-PCR products from three lymphoblastoid cell lines, three malignant cell lines, and four primary ALL cases were in accordance with the published cDNA sequences in each case. Polymorphic transcripts missing 11 bp at the beginning of exon 10 were detected in normal-sized cDNA as well as in aberrant products; the missing 11 bp do not affect the open reading frame. No correlation was detected between the occurrence of aberrant transcripts and the lack of Fhit protein, unlike results in solid tumors, where a good correlation was observed between the detection of aberrant FHIT RT-PCR products and the absence of the protein (9, 13).

We did not detect genomic rearrangements in any of the primary cases or cell lines by Southern blotting using a FHIT cDNA probe encompassing exons 3–10.

The FHIT gene at chromosome 3p14.2 is a tumor suppressor gene that is deleted or inactivated in a large variety of different human cancers. Deletion of both alleles is the most frequently observed event, resulting in the loss of function of the gene (27). Introduction of FHIT sequences into tumor cell lines suppressed their ability to form tumors in nude mice (19).

In this study, loss of Fhit protein expression was observed very frequently in primary ALL cases and leukemia/lymphoma cell lines. By Western blot analysis, only 24% of ALL cases or malignant cell lines expressed the Fhit protein, whereas all normal lymphoblastoid cell lines and PBL samples did. Loss of Fhit expression was found in all T-cell leukemias investigated. The low levels of Fhit protein detected in primary ALL cases after long exposure of the blots are probably derived from contaminating normal lymphocytes in the samples. Northern blotting of RNA from the cell lines confirmed the results obtained by Western blotting. Normal FHIT mRNA was detected in all normal lymphoblastoid cell lines but in only 2 of the 13 neoplastic cell lines investigated. These data strongly suggest that loss of Fhit expression occurs frequently in ALL samples, especially those derived from T-cells.

The results of two different RT-PCR approaches revealed a lack of normal FHIT transcripts in 33% of the malignant cell lines but in none of the normal lymphoblastoid cell lines and PBL samples. In the cell lines that did not show normal-sized cDNA by RT-PCR, Fhit protein or FHIT mRNA (by Northern blotting) was also not detectable.

Of the 18 primary ALL cases, 16 reproducibly showed normal-sized bands by RT-PCR. Because in 10 of these 16 cases only traces of Fhit protein were found by Western blot, it seems likely that the normal FHIT mRNA and protein may be derived from contaminating normal lymphocytes, detected more easily by the much more sensitive RT-PCR approach than by Western blot. Accordingly, in six of the leukemia/lymphoma-derived cell lines, normal-sized FHIT RT-PCR products were found, but neither Fhit protein nor FHIT mRNA transcripts (by Northern blotting) were detected. Thus, detection of the full-length FHIT RT-PCR product does not necessarily correlate with Fhit protein expression, because lack of normal transcripts in the main clone may be obscured by contaminating normal cells or by heterogeneity for the FHIT locus within the tumor clone itself (6, 10). These results indicate that in hematopoietic malignancies, RT-PCR may be too sensitive to detect loss of FHIT expression in the main clone, because even few interspersed normal cells will lead to a positive signal.

Fhit protein expression is low in normal lymphoid cells relative to epithelial tissues and cells, which is consistent with our failure to detect Fhit protein in lymphocytes using the immunocytochemistry approach. This result is also consistent with previous reports showing expression of Fhit in epithelial normal and tumor cells by immunohistochemistry, whereas infiltrating B and T cells appear negative or nearly negative for FHIT (13, 28). Other investigators (22, 23, 24, 25, 26) reported few primary leukemia cases lacking FHIT transcripts as detected by RT-PCR. Loss or down-regulation of Fhit protein expression was also found in a subset of primary renal cell carcinomas (4, 28), lung tumors (10), primary cervical carcinomas and cell lines (18), pancreatic carcinoma cell lines (14), and hematological malignancies (26). Together with our results, these data indicate that loss of FHIT expression may be involved in the pathogenesis and development of a large variety of human cancers, including ALL.

Aberrant bands were detected in nearly all tumor cell lines, as well as in all lymphoblastoid cell lines. They were also frequent in primary ALL cases and PBL samples, in concordance with results reported by Carapeti et al. (24) and Peters et al. (26), although in our cases, these products were mostly nonreproducible with regard to their occurrence and length. The exons lacking in the most prominent aberrant bands usually included exon 3, confirming the results of Sugimoto et al. (22), who found RT-PCR products missing exons 3–6 to be the main aberrant transcripts in various leukemias.

Sugimoto et al. (22) did not report aberrant RT-PCR products in their normal PBL samples, but other studies reported aberrant products from lymphoblastoid cell lines (29), EBV-immortalized B-cells (30), and PBLs (24, 26, 30, 31), similar to our results. Aberrant transcripts have also been reported in various other normal tissues as well as in the tissue of the corresponding tumor [liver (32), brain (33), cervix (34, 35), and prostate (36)]. This aberrant mRNA may reflect reduced splicing fidelity or may be due to rare rearrangement events occurring in the normal population similar to the t(14;18) and the t(9;22) chromosome translocations. These are observed specifically in follicular lymphoma and CML, respectively, but it is also possible to detect these rearrangements by PCR in PBLs of healthy individuals (37, 38).

Genomic alterations that would explain the loss of FHIT expression were not detected using a cDNA probe spanning exons 3–10 in Southern hybridization experiments. This is in concordance with the results of Sugimoto et al. (22) and Peters et al. (26), who did not find genomic alterations in their cases either. Because the FHIT locus spans a chromosomal region of over 1 Mb, overlapping deletions affecting exons of the gene may result in allelic alterations of FHIT without showing any genome rearrangement detectable by hybridization with cDNA (9).

Taken together, the lack of Fhit protein and FHIT mRNA in most of the ALL samples compared with the presence of the protein and mRNA in nearly all of the control samples indicates that the loss of Fhit expression may contribute to the pathogenesis of a considerable fraction of ALL cases. Our data suggest that even when DNA or alterations of the RT-PCR product in the protein coding region could not be observed, the Fhit protein and mRNA may be absent or dramatically reduced.

We thank Rebecca Connor for excellent technical assistance and Drs. Yuri Pekarsky and Raffaele Baffa for helpful discussions.