CML66, a broadly immunogenic tumor antigen, elicits a humoral immune response associated with remission of chronic myelogenous leukemia

Abstract

This report describes a tumor-associated antigen, termed CML66, initially cloned from a chronic myelogenous leukemia (CML) cDNA expression library. CML66 encodes a 583-aa protein with a molecular mass of 66 kDa and no significant homology to other known genes. CML66 gene is localized to human chromosome 8q23, but the function of this gene is unknown. CML66 is expressed in leukemias and a variety of solid tumor cell lines. When examined by Northern blot, expression in normal tissues was restricted to testis and heart, and no expression was found in hematopoietic tissues. When examined by quantitative reverse transcription–PCR, expression in CML cells was 1.5-fold higher than in normal peripheral blood mononuclear cells. The presence of CML66-specific antibody in patient serum was confirmed by Western blot and the development of high titer IgG antibody specific for CML66 correlated with immune induced remission of CML in a patient who received infusion of normal donor lymphocytes for treatment of relapse. CML66 antibody also was found in sera from 18–38% of patients with lung cancer, melanoma, and prostate cancer. These findings suggest that CML66 may be immunogenic in a wide variety of malignancies and may be a target for antigen-specific immunotherapy.

The therapeutic benefits of allogeneic bone marrow transplantation (BMT) derive in part from the antitumor effect of high-dose chemotherapy and radiation (1, 2). However, several clinical observations provide convincing evidence that donor immune elements also contribute to the elimination of residual leukemia after BMT. These observations include the reduced risk of relapse after BMT in patients who develop graft-versus-host disease and the increased risk of relapse in patients who receive T cell-depleted donor marrow (3, 4). It also has been demonstrated that relapse after BMT often can be successfully treated by donor lymphocyte infusion (DLI) without additional therapy (5–7). The demonstration that adoptive immunotherapy with donor T cells can provide long-lasting remissions provides compelling evidence that these cells play an important role in mediating a graft-versus-leukemia (GVL) response after allogeneic BMT (8, 9). Appreciation of the importance of GVL has led to the development of less intensive nonmyeloablative approaches for transplantation of allogeneic hematopoietic stem cells with subsequent infusion of donor T cells to enhance antitumor immunity (10–12). Initial reports using these approaches are encouraging and provide evidence that the therapeutic effects of DLI can be extended to provide effective immunity against solid tumors as well as hematopoietic malignancies (13).

Although reconstitution with allogeneic stem cells can provide effective antitumor immunity, the mechanisms whereby donor T cells exert this activity are unknown and the target antigens of this response have not been well defined. To better characterize the antitumor effect of DLI we previously examined the reconstitution of T and B cell immunity in patients with chronic myelocytic leukemia (CML) who received infusions of CD4⁺ donor lymphocytes for treatment of relapse after allogeneic BMT (14). Patients with CML were selected for this analysis because the great majority demonstrate a complete cytogenetic and molecular response within a defined time period after DLI and without additional intervention (15). These patients thus represent a unique opportunity to examine a consistently effective antitumor response in vivo. Although T cells are presumed to be the critical mediators of graft-versus-leukemia, our previous studies have shown that DLI initiates a complex immune response that includes a potent antibody response to a variety of leukemia-associated antigens (16). Using established methods for serological identification of tumor antigens by recombinant cDNA expression cloning (SEREX) (17, 18) we identified a panel of 13 leukemia-associated antigens that were recognized by high titer antibody 1 year after response to DLI. Within this panel of antigens, 11 represented known genes and two represented genes that had not previously been identified. This report describes the initial characterization of one of the additional genes, termed CML66.

Our characterization of CML66 demonstrates that this gene is highly expressed in different solid tumors and hematopoietic malignancies but high-level expression in normal tissues is restricted to testis and heart. Using a sensitive ELISA, we found that the highest titers of CML-specific IgG antibody were present in a patient with CML who responded to DLI. In this patient, the development of high titer specific antibody correlated well with the cytogenetic remission induced by DLI. Lower titers of specific antibody frequently were detected in other CML patients who responded to DLI. IgG antibodies specific for CML66 also were found in 18–38% of patients with lung cancer, melanoma, and prostate cancer. These observations indicate that CML66 antigen is broadly immunogenic in patients with different solid tumors as well as leukemia, and this response is not restricted to individuals who have undergone allogeneic marrow transplantation. The immunogenicity of this antigen and association with effective antitumor immunity in CML suggest that CML66 may be an appropriate target for immunotherapy in other malignancies.

Materials and Methods

Patient Samples.

Serum was obtained at various times before and after lymphocyte infusion in patients enrolled on a clinical trial of CD4⁺ DLI for treatment of relapse after allogeneic BMT (15). Serum samples also were obtained from patients with CML receiving therapy with hydroxyurea or IFN-α, patients with metastatic melanoma or metastatic nonsmall cell lung carcinoma enrolled into Institutional Review Board-approved tumor cell vaccine trials (19), and patients with hormone refractory advanced prostate cancer.

CML cDNA Library and Human Testis cDNA Library Screening.

CML cDNA library construction and screening have been described (16). Briefly, mRNA was extracted from peripheral blood mononuclear cells (PBMC) from three patients with CML by using standard methods and pooled to create a representational CML expression library in a λ bacteriophage expression vector. A human testis cDNA library (1 × 10⁶ phage) derived from normal whole human testes pooled from 11 males (CLONTECH) was screened with a 0.8-kb ³²P-labeled CML66 probe, as described (20). After three rounds of phage plaque purification, five positive clones were identified, converted into plasmid pTriplEx by cre-lox-mediated excision, and sequenced in both strands.

Northern Blotting.

Multiple tissue Northern blots were prepared with purified poly(A)⁺ RNAs (human cancer cell line blot, human normal tissue I blot, human normal tissue II blot, and human normal 12-lane blot, CLONTECH). Hybridizations were conducted with a 0.8-kb ³²P-labeled CML66 probe in the ExpressHyb hybridization solution (CLONTECH) at 68°C for 1 h according to the manufacturer's protocol. The same blots then were stripped and hybridized with the ³²P-labeled human β-actin cDNA probe as controls.

Human Genomic DNA Library Screening and Fluorescence in Situ Hybridization (FISH) Chromosome Localization Analysis.

Phage (1 × 10⁶) from a lambda Dash II human genomic DNA library (Stratagene) were screened by using described methods (20). Genomic DNA from purified positive phage were prepared by using Qiagen Lambda Midi Kit (Qiagen, Valencia, CA). The insert size of positive genomic DNA clones was determined by gel electrophoresis. Exon sequences in the genomic DNA clones encoding CML66 cDNA were confirmed by DNA sequencing.

Human FISH chromosome localization was performed by using a CML66 genomic clone with an insert of 23 kb labeled with digoxigenin dUTP by nick translation (Incyte Genomics, St. Louis). Labeled probe was combined with sheared human DNA and hybridized to metaphase chromosomes derived from phytohemagglutinin-stimulated peripheral blood lymphocytes in a solution containing 50% formamide, 10% dextran sulfate, and 2× SSC. Specific hybridization signals were detected by incubating the hybridized slides with fluorescein-conjugated antidigoxigenin antibodies followed by counterstaining with 4′,6-diamidino-2-phenylindole.

Reverse Transcriptase–PCR, PCR Cloning, and 5′ Rapid Amplification of cDNA Ends.

Total RNA was prepared from cultured tumor cell lines, patient CML cells, and normal human PBMC by using RNAzole (Tel-Test, Friendswood, TX). Reverse transcriptase–PCR and PCR cloning were performed as described (20). A sense primer (25k) specific for the 5′ upstream CML66 (5′-CGGAGAATTCGGCACGAGTCCCAGTCTCTGTGCGA-3′) and a second antisense primer (25c) specific for the 3′ downstream CML66 (5′-CGGAGAATTCTCATTCTCTGTATTTACTTTTATTAA-3′) were used for PCR cloning. All of the PCR cloning reactions were performed by using high-fidelity enzymes such as Pfu Turbo (Stratagene). The 5′ rapid amplification of cDNA ends by PCR was performed by using human testis Marathon-Ready cDNAs as templates with a CML66-specific antisense primer 25H (5′-CCCAGGTAGAAGATGAGAAATGGATA-3′) and the primer AP1 or AP2 specific for the adapter sequence (CLONTECH). PCR-amplified products were subcloned into the pCRII-TOPO vector (Invitrogen), followed by DNA sequencing.

Real-Time Quantitative PCR.

Quantitative PCR was performed by using Taq-Man PCR reagent kit on the ABI Prism 7700 Sequence Detection System (Perkin–Elmer) according to the manufacturer's protocol after reverse transcription of total RNA. Two PCR oligonucleotide primers, 25 forward (5′-TTAATTCTCACGCTGCGGC-3′) and 25 reverse (5′-GGTCTCTTCACCCGTAGGGAG-3′) were designed by using a PCR-Select program (Perkin–Elmer). A fluorescence-labeled oligonucleotide probe [5′-(6-Fam)TGGAAAGCGATGGAGGTGGCG(Tamra)-3′] specific for CML66 PCR product was used to monitor amplification of specific PCR product. Simultaneously, glyceraldehyde-3-phosphate dehydrogenase (GAPDH) copy number was measured in each cDNA sample as an internal control with a GAPDH amplification kit (Perkin–Elmer). GAPDH copy number varied only 10–20% among the samples. PCR conditions in three stages were as follows: stage 1, 50°C for 2 min; stage 2, 95°C for 10 min; stage 3, two-step cycling, 95°C for 15 sec, 60°C for 1 min for 50 cycles.

Generation of Glutathione S-Transferase (GST) Fusion Proteins.

A cDNA fragment encoding full-length long ORF (ORF 1) of CML66 with EcoRI restriction site on both ends was generated by PCR using high-fidelity enzyme Pfu Turbo DNA polymerase (Stratagene) and primers 25F1 (5′-CGGAGAATTCGATGGAGGTGGCGGCTAATTGCTCC-3′) and 25c (sequence described above). The underlined sequences in these primers were designed for subcloning into EcoRI site of GST fusion vector pGEX-3X (Amersham Pharmacia). All of these CML66 fragments were fused in-frame to the C terminus of GST protein after cloning into the EcoRI site of the GST expression vector pGEX-3X and were further examined by DNA sequencing before transformation into the BL-21 strain of the Escherichia coli. The GST and the full-length fusion protein GST-CML66 (25F1–25C, ORF1) were purified according to the manufacturer's protocols (Amersham Pharmacia) or with B-per Bacterial Protein Extraction Reagent (Pierce).

Western Blotting.

Recombinant proteins expressed in transformed E. coli were subjected to 10–12% SDS/PAGE with Tris-glycine buffer and transferred onto nitrocellulose filters in 20% methanol in Tris-glycine buffer. Proteins on the blots were visualized as described (16).

Detection of CML66-Specific Antibody in Patient Sera by ELISA.

ELISA plates (VWR Scientific) were coated with 50 μl of purified recombinant protein (GST or GST-CML66) at 5 μg/ml in coating buffer (PBS + 0.05% sodium azide) overnight at 4°C (16). Plates were washed with PBS with 0.05% Triton X-100 and blocked overnight at 4°C with 200 μl/well of 2% nonfat milk with 0.05% Triton X-100. A total of 50 μl/well patient sera was added to a final dilution of 1:200 and incubated at room temperature for 3 h. The procedure for detection of specific IgG antibody has been described (16). The specificity of selected positive samples also was confirmed by demonstrating inhibition of reactivity by preincubation of patient sera with purified recombinant CML66-GST.

Results

Expression of CML66 in Tumor Cell Lines and Normal Tissues.

In a previous report (16), we identified a novel 2.1-kb cDNA clone from a CML λ expression cDNA library. This clone was identified because of reactivity with sera obtained from a patient with CML who developed an effective immune response, resulting in complete remission of leukemia after donor lymphocyte infusion. The 2.1-kb clone had no significant sequence homology to any known genes in GenBank or other databases (National Center for Biotechnology Information, National Institutes of Health). As shown in Fig. 1A, Northern blot hybridizations with this cDNA probe showed that this gene had a 2.5-kb transcript and was highly expressed in seven of eight human tumor cell lines that were examined. These included HL-60, K562, Molt-4, and Raji cell lines derived from myeloid or lymphoid tumors as well as four cell lines derived from a variety of epithelial malignancies and melanoma. In contrast, Northern blots only revealed expression of CML66 in two of 26 normal human tissues (Fig. 1 B–D). As shown in Fig. 1 B and C, CML66 was expressed at relatively high levels in human testis and at lower levels in heart. Although increased binding was noted in pancreas, no specific hybridization band was noted in this tissue.

Figure 1.

Tissue expression of CML66 gene. Northern blots with poly(A) mRNA obtained from eight tumor cell lines and 28 normal tissues were hybridized with a CML66 cDNA probe. The size of the transcript is indicated on the left. β-actin mRNA loading controls for each lane were revealed by hybridization with a human β-actin cDNA probe.

Cloning of CML66 cDNA.

A normal human testis cDNA library was screened to clone the normal CML66 gene. The entire cDNA sequence was completed by using 5′ rapid amplification of cDNA ends. These experiments identified a 2,319-bp sequence that contains 242 bp of 5′ untranslated region (UTR), a 1,749-bp ORF with 583 aa, and a 338-bp 3′ UTR (based on computer analysis). The DNA sequence at the start codon in the ORF contained a Kozak consensus sequence (A/GNNATGG) for high-efficiency protein translation (21). In vitro transcription and translation confirmed that this long ORF encoded a 66-kDa protein (not shown). A polyadenylation signal (AATAAA) was found in the 3′ UTR. In addition, 5′ end primer extension experiments (Promega) indicated that the transcription starting site was located 200 bp upstream of the 5′ end of this cloned transcript. This correlated well with the 2.5-kb size of the gene shown in Northern blots. Because this gene product was 66 kDa in size and originally was isolated from a CML library, it was termed CML66.

cDNA Sequence Comparison of CML66 Gene in Normal Tissues and Tumor Cells.

CML66 cDNA was cloned by screening a normal human testis cDNA library using CML66 cDNA isolated from the CML library as a probe. Five separate clones of different lengths were sequenced in both strands, and all overlapping regions were found to have identical sequence. Comparison of the normal CML66 gene isolated from the testis library with the sequence isolated from the CML library demonstrated that the cDNA sequences were identical except for two single nucleotide differences (Table 1). One substitution at bp 1412 resulted in a change from Asn to His at amino acid 394. A second substitution at bp 1509 resulted in a change from Asn to Ser at amino acid 426. Three cDNA clones of different lengths isolated from the CML library were sequenced in both strands, and all contained these two single nucleotide differences.

Table 1.

Single nucleotide differences in CML66 cDNA amplified from tumor cells compared to normal CML66 cDNA from human testis

Changes		CML						AML P1	Jurkat	A549	MCF7
Nucleotide	Amino acid	Library	P1	P2	P3	K562	TF1
C243T	A4V			+
G512A	A94T			+
T440G											+
A614G	T128A							+
A692G	I154V						+
C756T	A175V								+
A903G	D224G								+
C988A	F252L			+	+	+			+
T996C	V255A								+
A1008G											+
G1037A	V269I			+	+	+			+		+
T1148C	F306L			+
A1308G	E359G				+
A1412C	N394H	+	+
A1509G	N426S	+	+
T1566C	V445A			+
T1696C							+
A1832G	K534E					+
T1857C	V542A						+

CML66 cDNA sequence was determined in a CML library, cDNA from four patients with leukemia (CML P1–3 and AML P1) and five cell tumor cell lines. Nucleotide changes are designated as T440G (for example), where T indicates the nucleotide at position 440 in normal testis and G indicates the nucleotide at the same position in the tumor cell source. Amino acid changes are designated in a similar fashion. 

CML66 cDNA was amplified by high-fidelity PCR from leukemia cells from three additional patients with CML, one patient with acute myelogenous leukemia and a panel of tumor cell lines. CML cells from one of these patients (CML-P1) had been used to construct the CML cDNA library, and the CML66 sequence in this individual was identical to the CML library sequence. DNA sequence of nine CML66 clones from these tumor cells was compared with the sequence derived from normal testis and 17 additional single-bp mutations were identified (Table 1). Three mutations were silent but 14 mutations resulted in amino acid substitutions. None of these mutations resulted in premature stop codons or reading frame shifts and two mutations (F252L and V269I) occurred in multiple tumor cells.

Quantitative Analysis of CML66 Expression in Normal PBMC and CML.

Because Northern blots did not demonstrate significant expression of CML66 in normal PBMC, we developed a more sensitive quantitative PCR assay to detect CML66 gene expression and compare expression in primary CML cells with normal PBMC. As shown in Fig. 2, CML66 copies were significantly increased in five CML samples compared with normal PBMC but this increase was only ≈1.5-fold. For comparison, CML66 copies in tumor cell lines positive by Northern blot were 30-fold higher than in normal PBMC (data not shown).

Figure 2.

Relative expression of CML66 gene in normal PBMC and primary CML. CML66 mRNA copy number was determined in total RNA from PBMC from normal donors (n = 5) and primary peripheral blood CML cells (n = 5). CML66 copy number measured by real-time quantitative PCR was significantly increased in CML compared with normal PBMC (P = 0.002).

Chromosome Localization of CML66.

Restriction enzyme analysis of normal human genomic DNA followed by Southern blot hybridization with CML66 cDNA probe suggested that CML66 was a single copy gene (data not shown) (22). Human chromosome localization of CML66 was performed by FISH using a 23-kb CML66 genomic DNA clone as a probe. A total of 80 metaphase cells were analyzed with 62 (78%) exhibiting specific labeling. Based on size and morphology of specifically labeled chromosomes, CML66 was localized to chromosome 8. Cohybridization with CML66 clone and an anonymous genomic clone known to map to 8q12 resulted in labeling of the long arm of chromosome 8 at two distinct loci. Measurement of 10 specifically labeled chromosome 8 demonstrated that CML66 is located at a position 67% of the distance from the centromere to the telomere of chromosome arm 8q, an area that corresponds to band 8q23.3 (data not shown) (23).

Antibody Response to CML66 After DLI.

To characterize the immunogenicity of CML66 as a tumor rejection antigen, CML66-GST fusion protein was purified and used as a probe to analyze antibody reactivity in normal and CML patient sera. Purified CML66-GST fusion protein has a molecular mass of 96 kDa corresponding to the combined size of GST (30 kDa) plus the complete ORF of CML66. In the Western blot shown in Fig. 3, antibodies to CML66 were not detected in normal sera but were detected in sera obtained from a patient with CML 1 year after DLI. Antibodies to GST-CML66 were not detected in patient serum obtained before allogeneic BMT or before DLI. Serum from this patient had been used to screen the CML library and this result therefore confirmed that the CML66 protein had been immunogenic in vivo.

Figure 3.

Analysis of CML66 immune reactivity by Western blot. Purified GST-CML66 fusion protein was loaded onto separate lanes as indicated. Western blots were probed with anti-GST antibody, revealing a 96-kDa band in all lanes containing GST-CML66 fusion protein (Upper). A 30-kDa protein was detected in control lanes loaded with purified GST only (not shown). This Western blot subsequently was probed with CML patient sera (Lower). Reactivity of CML patient sera collected pre-BMT (Pre-B), pre-DLI (Pre-D), and post-DLI (Post-D) was compared with normal donor sera. GST-CML66 (96 kDa) was detected only with post-DLI serum.

To provide a more sensitive method for detecting and quantifying the immune response to CML66 we developed an ELISA using purified CML66-GST. IgG antibodies to CML66 were not detectable by ELISA before BMT and before DLI, but antibody titers increased markedly 3 months post-DLI and persisted at high levels for at least 1 year (Fig. 4). Specific antibody was no longer detectable 5 years after DLI. The time course of antibody reactivity in this patient correlated well with the onset of cytogenetic response. After achieving a complete cytogenetic remission 3 months post-DLI, BCR-ABL mRNA remained detectable in blood and bone marrow until a molecular remission was achieved 12 months post-DLI. Further characterization of the antibodies reacting with CML66 demonstrated that they were primarily IgG1 and IgG4 isotypes (data not shown).

Figure 4.

Correlation of anti-CML66 IgG with cytogenetic response after donor lymphocyte infusion. Quantitative assessment of CML66-specific IgG antibody was determined in serial serum samples from a patient with relapsed CML who responded to DLI. The percent marrow metaphases containing the Philadelphia chromosome as well as results of PCR analysis of patient samples for the presence of BCR-ABL mRNA also are indicated.

Quantitation of IgG Response to CML66 in Normal Donors and Patients with Cancer.

The CML66 ELISA also was used to determine levels of specific IgG antibody in sera obtained from normal donors and patients with different malignancies. In this assay, results are expressed as a ratio of reactivity with purified CML66-GST compared with reactivity with GST alone. As summarized in Fig. 5, reactivity was not detected in sera from normal donors (n = 12) but specific CML66 reactivity was detected in one of 15 patients with CML receiving hydroxyurea or IFN-α. Reactivity with CML66 was present in seven of 13 patients with relapsed CML after allogeneic BMT who had responded to donor lymphocyte infusion. CML66 reactivity also was found in patients with lung cancer (four of 13 patients), melanoma (four of 22 patients), and prostate cancer (15 of 39 patients). The highest level of reactivity was observed in the patient with CML known to have specific antibody by Western blot (Fig. 3). These results indicate that CML66 is capable of eliciting a humoral immune response in patients with a variety of solid tumors.

Figure 5.

Quantitative anti-CML66 IgG measured by using ELISA in serum from normal donors and patients with CML, lung cancer, melanoma, and prostate cancer. Serum samples from 12 normal donors, 15 patients with CML receiving conventional therapy, 13 with CML after response to DLI, 13 with lung cancer, 22 with melanoma, and 39 with prostate cancer were tested at a dilution of 1:200. Results are expressed as a ratio of CML66-GST reactivity compared with reactivity with GST alone. The dashed line represents the upper limit of reactivity in serum from 12 normal donors (mean ± 2 SD = 1.52).

Discussion

In patients with relapsed CML after allogeneic BMT, DLI initiates an effective antitumor response that results in the elimination of leukemia cells in more than 70% of patients (5, 6, 15). Almost all patients who achieve a cytogenetic response also subsequently become PCR-negative for cells containing BCR-ABL transcripts. Thus far, few patients have relapsed after achieving a complete molecular response. These clinical observations demonstrate that the antileukemia response associated with DLI results in the elimination of relatively large numbers of tumor cells in vivo and the development of long-lasting tumor immunity. However, despite the effectiveness of DLI, the target antigens of this immune response have not been well characterized. In particular, it is not known whether the immune response is primarily directed at leukemia-specific antigens, allogeneic antigens, or other tumor-associated antigens. To characterize the antigenic targets of the DLI response we previously demonstrated the presence of high titer antibody responses to a variety of proteins expressed by leukemia cells (16). The present studies focus on a single antigen that was identified in these experiments, and our characterization of this immunogenic target provides insights into the complex immune response that is initiated in these individuals.

After establishing the cDNA sequence of the CML66 gene, we examined the expression of this gene in a variety of human tumor cell lines and normal tissues. Northern blots demonstrated abundant mRNA in most tumor cell lines, but the expression of CML66 in normal tissues was very restricted. High-level expression appeared to be limited to testis and lower-level expression was found in heart. Normal hematopoietic tissues including bone marrow, peripheral blood lymphocytes, spleen, and thymus did not have detectable gene expression by Northern blot. To compare levels of CML66 gene expression in normal and malignant hematopoietic cells, we therefore developed a sensitive quantitative reverse transcriptase–PCR assay and found that primary CML cells expressed significantly higher levels of CML66 transcripts than normal PBMC. However, this represents only an approximately a 1.5-fold increase in mRNA copy number, and we do not yet know to what extent CML66 protein expression is increased in these cells. Although the CML66 gene is not homologous to other cancer-testis antigens, this gene appears to be widely expressed in hematopoietic and nonhematopoietic tumors with only limited expression in normal tissues.

The CML66 gene was found to encode a protein with molecular mass of 66 kDa. The cDNA sequence has not previously been reported and there are no significant homologies with other known genes in GenBank. CML66 cDNA sequence contains no defined functional motifs and thus the function of this gene is entirely unknown. We have established that CML66 gene is localized to human chromosome 8q23. The National Cancer Institute database (Cancer Genome Anatomy Project) shows that balanced and unbalanced chromosomal abnormalities involving chromosome 8q23 have been found in some leukemias and non-Hodgkin's lymphomas. The 8q arm is also frequently amplified in advanced prostate cancer (24, 25). However, sequence analysis of CML66 in several tumor cell lines did not demonstrate any insertions or deletions indicating direct involvement of a translocation. Further studies will be necessary to determine the mechanism for increased expression of CML66 in human tumors and to identify the function of this gene.

To investigate whether genetic mutations or polymorphisms might contribute to the immunogenicity of CML66, cDNA was amplified from nine different tumor cell populations, and each DNA sequence was compared with the normal sequence cloned from a human testis library. Nucleotide substitutions resulting in amino acid differences were found in eight of the nine tumor cells we examined. Most tumor cells had multiple substitutions, and two identical substitutions each were found in four and five tumor populations, respectively. Some of these substitutions might reflect either PCR or sequencing errors but the finding of identical substitutions in different tumor cells make this possibility less likely. In addition, five different clones of distinct size were sequenced from the testis library. Although this library was constructed with tissue from 11 different males, no nucleotide differences were found in these different clones. This suggests that the CML66 gene is not highly polymorphic, but further analysis of CML66 sequence in both normal tissues and tumor cells from the same individuals will be necessary to determine which of these nucleotide substitutions reflect mutational events in the tumors and which might reflect genetic polymorphisms.

We also directly examined whether the CML66 gene product was the target of a specific antibody response in vivo. Full-length CML66-GST fusion protein was expressed and purified, and patient serum was obtained after response to DLI was reactive with this fusion protein by Western blot. These observations confirmed that the CML66 gene was immunogenic in vivo and an ELISA using purified GST-CML66 was established to quantify this response. Using this sensitive ELISA, we confirmed that antibodies specific for CML66 were not detectable before allogeneic BMT or before DLI and found that the response to CML66 coincided with the elimination of CML cells in vivo. Although this patient demonstrated an effective antileukemia response, she did not develop clinically significant acute or chronic graft-versus-host disease. High titer-specific antibody persisted for at least 1 year and this patient remains in complete cytogenetic and molecular remission 5 years after DLI. Despite the expression of CML66 in heart, this patient has had no evidence of cardiac toxicity after DLI.

The persistence of high titer IgG for prolonged periods suggests that the CML66 protein is the target of a coordinated immune response involving both B and T cells. Other tumor-associated antigens such as NY-ESO-1 were first identified by using a SEREX approach and subsequently were shown to be targets of specific T cell immunity in those individuals who had developed specific B cell responses (26–28). Similarly, antigens such as MAGE that were first identified by cytotoxic T cells have been subsequently shown to also be targets of antibody responses in vivo (29). Having demonstrated that CML66 is the target of an antibody response after DLI, further studies can be directed to determine whether CML66 epitopes also are recognized by specific T cells in the same patients. Taken together, these studies will help define the role of T and B cell immunity to CML66 in the graft-versus-leukemia response.

The observation that CML66 was highly expressed in epithelial tumor cell lines as well as leukemias prompted us to examine whether antibodies to CML66 could be detected in patients with different solid tumors. Using ELISA, we detected specific IgG reactivity to CML66 in patients with lung cancer, melanoma, and prostate cancer. CML66-specific antibodies were found in 18–38% of patients with these solid tumors. All of these patients had relatively advanced disease and several were enrolled in clinical trials evaluating tumor vaccines (19). Further follow-up of those patients who received tumor vaccines will be helpful to determine whether the response to CML66 was enhanced by vaccination and whether the response to this antigen correlated with tumor regression. Moreover, ELISA also can be used to screen sera from larger numbers of patients and to determine whether the development of immunity to CML66 may define distinct populations of patients with these cancers.

Patients who respond to DLI develop a complex and coordinated response to a diverse set of antigenic targets (16). This potent response often can be achieved without clinical evidence of graft-versus-host disease, suggesting that the targets of this immune response are not widely expressed by normal recipient cells. Our analysis of the expression of CML66 and the antibody response to this gene suggests that the DLI response is, in part, directed against this tumor-associated antigen. Although CML66 expression is not restricted to leukemia cells, the expression of this gene in normal tissues is limited, which may contribute to the lack of toxicity associated with the DLI response (15). In patients with solid tumors who undergo allogeneic stem cell transplantation, immunity to antigens like CML66 may help explain the effectiveness of the antitumor immune response (13, 30). Although further studies will be needed to determine the clinical significance of the humoral response to CML66, our initial observations suggest that relatively large numbers of patients with different solid tumors develop specific immunity to this antigen. Considering the effectiveness of DLI and lack of toxicity associated with this immune response, CML66 may provide a target for immunotherapeutic approaches in a variety of solid tumors.

Abbreviations

CML

chronic myelogenous leukemia

BMT

bone marrow transplantation

DLI

donor lymphocyte infusion

PBMC

peripheral blood mononuclear cells

FISH

fluorescence in situ hybridization

GST

glutathione S-transferase