The study was performed at the Veterinary Medical Center of the University of Tokyo. The study was presented at 29th Annual Conference of Veterinary Cancer Society in 2009.
Increase in Minimal Residual Disease in Peripheral Blood before Clinical Relapse in Dogs with Lymphoma that Achieved Complete Remission after Chemotherapy
Article first published online: 11 FEB 2011
Copyright © 2011 by the American College of Veterinary Internal Medicine
Journal of Veterinary Internal Medicine
Volume 25, Issue 2, pages 292–296, March/April 2011
How to Cite
Sato, M., Yamazaki, J., Goto-Koshino, Y., Takahashi, M., Fujino, Y., Ohno, K. and Tsujimoto, H. (2011), Increase in Minimal Residual Disease in Peripheral Blood before Clinical Relapse in Dogs with Lymphoma that Achieved Complete Remission after Chemotherapy. Journal of Veterinary Internal Medicine, 25: 292–296. doi: 10.1111/j.1939-1676.2010.0675.x
- Issue published online: 7 MAR 2011
- Article first published online: 11 FEB 2011
- Submitted August 26, 2010; Revised November 24, 2010; Accepted December 13, 2010.
- Lymphoid tumor;
Background: We developed previously a minimal residual disease (MRD) monitoring system in dogs with lymphoma by exploring a highly sensitive real-time PCR system.
Objectives: To identify the change in MRD before clinical relapse in dogs with lymphoma that achieved complete remission after chemotherapy.
Animals: Twenty dogs with multicentric high-grade B-cell lymphoma.
Methods: MRD levels in peripheral blood mononuclear cells (PBMCs) were measured by real-time PCR amplifying the rearranged immunoglobulin heavy chain gene. MRD measurement and clinical assessment were performed every 2–4 weeks for 28–601 days after completion of chemotherapy. An increase in MRD was defined as an increase by more than 0.5, calculated by log10[copy number of MRD per 105 PBMCs], based on the uncertainty level observed in a canine lymphoma cell line.
Results: During the follow-up period, 15 dogs relapsed in 28–320 days (median, 120 days) after completion of chemotherapy. An increase in MRD was detected 2 weeks or more before relapse in 14 of the 15 dogs, but an increase in MRD before relapse could not be detected in the remaining 1 dog. The time from increased MRD to clinical relapse was 0–63 days (median, 42 days). In contrast, no increase in MRD was detected in 5 dogs that did not experience clinical relapse.
Conclusion and Clinical Importance: An increase in MRD can be detected before clinical relapse in dogs with lymphoma. Application of early reinduction therapy based on an increase in MRD before clinical relapse may improve treatment outcome in canine lymphoma.
diffuse large B-cell lymphoma
fine needle aspiration
immunoglobulin heavy chain
minimal residual disease
peripheral blood mononuclear cells
a 6-month modified version of the University of Wisconsin–Madison chemotherapy protocol
Canine lymphoma is one of the most chemoresponsive malignancies in dogs. Initial response rates with chemotherapy are reported to be as high as 69–94%.1–4 However, nearly all treated dogs experience relapse and die because of disease progression. Despite the fact that various multidrug chemotherapeutic protocols have been introduced, the median survival time in a recent study was approximately 12 months.5 In general, the treatment of relapsed lymphoma results in a lower response rate and shorter remission duration compared with untreated lymphoma.6–8 Early detection of tumor relapse and prompt therapy are a reasonable approach for managing recrudescent disease.
Minimal residual disease (MRD) is defined as residual malignant cells that escape anticancer therapies and which are considered to be the source of tumor relapse.9,10 Monitoring MRD is an essential tool for risk group stratification and appears to be useful for predicting relapse in some hematopoietic and lymphoid malignancies of humans.10,11–13 In addition, several reports document the usefulness of MRD to provide a molecular assessment of treatment efficacy for diffuse large B-cell lymphoma (DLBCL) in humans,14–16 which is the disease to which high-grade B-cell lymphoma in dogs has been compared. We previously reported a highly sensitive MRD detection system in canine lymphoma by a real-time PCR that enabled detection of 1 tumor cell per 104 peripheral blood mononuclear cells (PBMCs).17 Our recent study indicated that MRD at the end of chemotherapy was correlated with outcome in dogs with lymphoma.18
Real-time PCR-based monitoring is considered useful for detecting early changes in tumor cell growth in the body before clinical relapse. In the present study, we carried out prospective PCR monitoring of tumor cell numbers in the PBMCs of 20 dogs with lymphoma that had achieved complete remission (CR) after chemotherapy, most of which subsequently experienced clinical relapse during the observation period.
Materials and Methods
Between April 2006 and August 2009, 20 dogs diagnosed cytologically with high-grade B-cell lymphoma by fine needle aspiration (FNA) of enlarged peripheral lymph nodes according to the updated Kiel classification19,20 were enrolled in this study. B-cell lineage was determined by the detection of clonal immunoglobulin heavy chain (IgH) gene rearrangement by PCR for antigen receptor gene rearrangement with primers described in 3 previous reports.21–23
This study was performed at the Veterinary Medical Center of the University of Tokyo and 3 private animal hospitals. Cytologic diagnosis and classification were agreed upon by two of the authors of this study (M.S. and Y.F.). Further eligibility criteria were as follows: (1) anatomic form of lymphoma was multicentric, (2) clone-specific oligonucleotides primers and probes could be successfully designed for measuring MRD, and (3) dogs achieved CR and completed a 6-month modified version of the University of Wisconsin—Madison chemotherapy protocol (UW-25).2
Pretreatment evaluation for all dogs included physical examination, CBC, serum biochemistry profile, thoracic and abdominal radiographs, and abdominal ultrasound examination. Clinical staging was performed according to the World Health Organization criteria for canine lymphoma24 except for the assessment of bone marrow involvement. Classification as stage V was determined by detection of neoplastic lymphoid cells in circulation by light microscopy.
Follow-Up Evaluation and Sampling for MRD Measurement
Response evaluation was performed according to the Veterinary Cooperative Oncology Group consensus document25 with a cytologic finding of lymphoma on FNA. Routine follow-up evaluation was conducted every 4 weeks after the completion of the chemotherapy protocols. Follow-up evaluation was conducted more frequently (every 2 weeks) after an increase in MRD was detected until the identification of progressive disease. Blood samples were obtained at diagnosis, postchemotherapy (2 weeks after the last injection of chemotherapeutic agent), and each follow-up evaluation.
MRD Measurement with Real-Time PCR
MRD measurement was performed as described previously.17 Briefly, to determine the sequence of the rearranged antigen receptor gene in each lymphoma case, the PCR products amplified from the rearranged IgH gene fragments were ligated into a T/A cloning vectora and subjected to sequence analysis. In order to obtain the sequence of the 3′-flanking region of the rearranged IgH J gene segments, their sequences were searched in the dog genome database by the BLAST search program (National Center for Biotechnology Information). Allele-specific oligonucleotides complementary to the complementarity-determining region 3 of the IgH genes of each dog were designed by a primer-designing software program.b
To assess MRD in PBMCs, peripheral blood samples (3 mL) were collected in EDTA-treated tubes, followed by density gradient centrifugation with Ficoll/Hypaque (specific gravity, 1.077).c Blood samples collected at private animal hospitals were shipped to the University of Tokyo in a container kept at 4°C, and PBMCs were separated within 24 hours of collection. The genomic DNA of the PBMCs was extracted with a DNA extraction kit.d Concentration of the extracted genomic DNA was calculated based on absorbance at 260 nm measured with a spectrophotometer.
Real-time PCR was performed with a thermal cycler,e and the data were analyzed by the software supplied by the manufacturer. To normalize the amount of DNA in samples, the albumin gene was used as an internal reference.17
Two consecutive samples (a new sample to be evaluated and a sample with known MRD from the last examination) were evaluated simultaneously. The accuracy of the calculated MRD was assessed from the interrun variability of the results from a control DNA sample derived from a canine lymphoma cell line.17 All measurements were conducted in triplicate.
Measurement Uncertainty Level for MRD Detection with Real-Time PCR
In order to determine the measurement uncertainty level of this assay, template DNAs derived from a canine lymphoma cell line17 corresponding to 10, 102, 103, 104, and 105 copies (the measurable range by this method) of the rearranged gene were subjected to real-time PCR. Six samples of each number of copies were subjected to real-time PCR to calculate the mean and standard deviation (SD) of the copy number of the tumor cells. The copy number of the tumor cells was transformed into log10. As the number of reference lymphoma cells became smaller in this assay, the SD became larger. When 10 reference lymphoma cells were included in the template, the resultant SD was largest (0.22) as calculated by the formula log10[copy number of tumor cells]. Thus, we tentatively defined an increase in MRD as an increase of more than 2 SD (0.44). Regarding the results of the control lymphoma cell line, if MRD increased by >0.5 (calculated from log10[copy number of MRD per 105 PBMCs]) in comparison with that from the previous visit, we judged this as an increase in MRD.
The number of tumor cells at diagnosis and subsequent MRD at postchemotherapy, admissions for follow-up, and clinical relapse were analyzed by the Krsukal-Wallis test to assess for differences. Differences of MRD between the postchemotherapy and other time points were evaluated by Dunnett's test. All statistics were calculated by statistical software.f Values of P < .05 were considered significant.
The 20 dogs in this study belonged to 13 breeds: the most commonly represented were Welsh Corgi (3) and Miniature Schnauzer (3) followed by Pug (2), Miniature Dachshund (2), mixed (2), Standard Poodle (1), Bulldog (1), Golden Retriever (1), Shih Tzu (1), Australian Kelpie (1), Polish Lowland Sheepdog (1), Yorkshire Terrier (1), and French Bulldog (1). Median age was 7 years (range, 2–13 years) and median body weight was 10.3 kg (range, 3.0–36.3 kg). Ten dogs were male (3 were castrated) and 10 were female (6 were spayed).
At diagnosis, 4 dogs (20%) were classified as clinical stage III, 10 (50%) as IV, and 6 (30%) as V. Six dogs (30%) were classified as substage b and 14 (70%) as substage a.
Eighteen of the 20 dogs did not receive any treatment for lymphoma before entering this study; these were treated with the UW-252 and achieved CR. Two dogs had received the 1st round of UW-25 and achieved CR. The 2 dogs experienced relapse 3–4 months later. After receiving a 2nd round of UW-25, both cases achieved CR again and were enrolled in the present study.
Median follow-up duration after chemotherapy was 133 days (range, 28–601; 95% confidence interval [CI], 104–235). Fifteen of the 20 dogs experienced relapses during the observation period. The median duration from the end of chemotherapy to relapse in these 15 dogs was 120 days (range, 28–350; 95% CI, 82.5–203). The other 5 dogs did not experience relapse during follow-up (92, 119, 173, 374, and 601 days).
Number of Tumor Cells in PBMCs at Diagnosis and Change in MRD in 15 Dogs that Experienced Clinical Relapse during Follow-Up
In the 15 dogs that experienced clinical relapse during follow-up, MRD was measured at several points before and at the time of clinical relapse (Fig 1).
The number of tumor cells indicated as log10[copy number of tumor cells per 105 PBMCs] at diagnosis ranged from 3.18 to 5.0 (median, 4.58) in the 15 dogs (Fig 1). Postchemotherapy (2 weeks after the last injection of chemotherapeutic agent) MRD decreased to <1.0–1.83 (median, 1.19). In 6 of the 15 dogs, MRD was lower than the detection limit of the real-time PCR assay used in this study. A significant difference was observed between the number of tumor cells at diagnosis and MRD postchemotherapy (P < .001) (Fig 1).
At clinical relapse, MRD was increased to 2.82–4.97 (median, 3.29), a significant increase compared with MRD postchemotherapy (P < .001) (Fig 1).
In these 15 cases, MRD was evaluated at 2–4 week intervals before identification of clinical relapse. MRD 4 weeks before clinical relapse ranged from <1.0 to 3.17 (median, 2.19). MRD 2 weeks before clinical relapse (range, 2.2–3.68; median, 2.67) was significantly greater than MRD postchemotherapy (P < .01) (Fig 1).
Time Points of Increased MRD in Dogs that Experienced Clinical Relapse
During follow-up in CR, both monitoring MRD and clinical assessment by physical examination, CBC, and diagnostic imaging were carried out every 4 weeks during the initial phase. However, when MRD increased by >0.5 in comparison with the previous admission, the interval of evaluation was shortened to 2 weeks.
In one case, MRD was 1.21 postchemotherapy; after 4 weeks (6 weeks before clinical relapse), it increased to 1.74. MRD in this case progressively increased to 1.95 and 2.22 after 2 and 4 weeks, respectively. At these time points, clinical assessment indicated the dog was still in CR. However, relapse was detected 6 weeks later, and MRD at this point was 3.47. Changes in MRD indicated that an increase in MRD of >0.5 could be detected 6 weeks before clinical relapse. Similar results were obtained in the other 13 cases.
In contrast, an increase in MRD before relapse could not be detected in 1 dog. This dog had postchemotherapy MRD <1.0; but it exhibited apparent clinical relapse 4 weeks later with a distinct increase in MRD (2.82).
The median duration from the time at which MRD increased to clinical relapse in the 15 dogs was 42 days (range, 0–63; 95% CI, 33–50.5). When the increase in MRD was defined as an increase of >0.5, the sensitivity for predicting relapse using an increase in MRD in consecutive samples obtained at 4-week intervals was 93.3% (14/15).
MRD in Dogs that Did not Experience Clinical Relapse during the Follow-Up Duration
In 5 of the 20 dogs, no clinical relapse was detected during follow-up. MRD in these dogs was evaluated at 4-week intervals during the follow-up period.
In 1 case, postchemotherapy MRD decreased to below the detection limit (1.0) of the assay. MRD 4, 8, 12, and 16 weeks postchemotherapy remained low (<1.0 in this case). Similarly, an increase in MRD of >0.5 was not observed during the follow-up period (92–601 days) in the other 4 dogs. The specificity for predicting relapse using an increase in MRD (>0.5) was 100% (5/5) in the consecutive samples obtained at 4-week intervals.
Based on previous studies describing qualitative26 and quantitative PCR17,18 assays to detect neoplastic lymphoid cells in the peripheral blood of dogs both before chemotherapy and after achieving CR, we decided to examine the clinical usefulness of MRD monitoring to predict clinical relapse in dogs with lymphoma. The number of tumor cells and MRD in peripheral blood were correlated with the clinical status of patients. There were statistically significant differences in MRD among the samples at the following points: diagnosis and postchemotherapy, postchemotherapy and clinical relapse, and postchemotherapy and 2 weeks before clinical relapse.
In the present study, an increase in MRD was tentatively defined as an increase of >0.5 as calculated from the value of log10[copy number of MRD per 105 PBMCs], based on the measured uncertainty level obtained from the data generated using a canine lymphoma cell line. This threshold value appeared appropriate because it provided a high sensitivity (93.3%) and specificity (100%) for predicting relapse in canine lymphoma. Moreover, we could infer tumor cell growth from the increase in MRD 42 days (median) before actual clinical relapse. Therefore, MRD monitoring of peripheral blood after the completion of chemotherapy is useful for detecting early growth of tumor cells that eventually leads to clinical relapse. One dog relapsed without a preceding increase in MRD 4 weeks before. At the time of apparent clinical relapse, this dog experienced an increase in MRD. In this case, it is conceivable that progression from CR to clinical relapse was rapid. Consequently, we could not detect the increase in MRD before relapse. It is possible that consecutive MRD monitoring at 4-week intervals could miss rapid tumor cell growth leading to clinical relapse. Therefore, more frequent monitoring may be required to detect rapid tumor cell growth. However, such cases are considered to be infrequent in dogs with B-cell high-grade lymphoma.
We judged the presence of an increase in MRD from the relative increase compared with the value 4 weeks before for predicting clinical relapse. Alternatively, an absolute value of MRD also can be helpful for detecting tumor cell growth leading to clinical relapse. When we found MRD increasing over the absolute value of 1.5, we could detect the tumor cell growth in 11 of the 15 dogs (73.3%) that experienced clinical relapse in our case series. Using this criterion, duration from the point of MRD increase to clinical relapse was 0–63 days (median, 37 days). When it was defined as an increase >2.0, tumor cell growth before clinical relapse could be detected in 14 of the 15 dogs (93.3%) that experienced clinical relapse. However, duration from the point of the increase in MRD to clinical relapse was shortened to 0–43 days (median, 28 days). Therefore, in terms of the sensitivity and time until clinical relapse, an MRD increase defined as a relative increase >0.5 in comparison with the value at the previous visit would be more useful to predict subsequent clinical relapse.
In all of the 15 dogs that exhibited progression of relapse in this study, the real-time PCR system at diagnosis was able to amplify the IgH gene of the lymphoma cells proliferating before and during clinical relapse. An increase in MRD correlated with the progression from CR to relapse in these cases. These results indicate that the lymphoma cells proliferating at relapse were derived from the same clones that formed the initial lymphoma lesions at diagnosis. Therefore, it is conceivable that clonal change of the tumor cells is an infrequent phenomenon in canine B-cell high-grade lymphoma.
In hematopoietic and lymphoid malignancies in humans, recent studies reported that some patients with increasing MRD were successfully treated with antineoplastic agents to avoid overt relapse.27,28 Preemptive therapy is shown to be effective for preventing morphologic or cytogenetic relapse after molecular relapse in acute promyelocytic leukemia.27 In another report, rituximab induced the effective clearance of MRD detected in molecular relapses of mantle cell lymphoma, resulting in long remission durations.28 Studies on MRD with clinical relevance in humans have been reported predominantly in mantle cell lymphoma and follicular lymphoma in which disease-specific genetic markers derived from chromosomal translocation are applicable in many patients. In contrast, in human DLBCL, such a genetic marker cannot be used. Hence, to set up an MRD monitoring system in human DLBCL, preparation of patient-specific IgH gene primers is needed as described in the present study in dogs. For these reasons, there have been only a small number of studies on MRD monitoring in human DLBCL.14–16 Therefore, we believe our study in dogs can provide beneficial information for human DLBCL.
In conclusion, we successfully detected an increase in MRD before relapse in canine high-grade lymphoma, indicating that molecular relapse is a harbinger of impending clinical relapse. The present study provides a basis for conducting a novel therapeutic strategy of early reinduction therapy based on detection of increased MRD in dogs with lymphoma.
a pGEM-T Easy, Promega Corporation, Leiden, the Netherlands
b Primer Express software program version 2.0, Applied Biosystems, Foster City, CA
c Lymphoprep, Nycomed Pharma AS, Oslo, Norway
d QIAamp Blood mini kit, Qiagen, Vallencia, CA
e Takara TP800, Takara Bio Inc, Tokyo, Japan
f Statmate3, ATMS, Tokyo, Japan
We thank Drs Kaori Takahashi, Keiichi Fujita, Masashi Yuki, and Jun Kinoshita for providing samples from dogs with lymphoma for this study.
This work was supported by a Grant-in-Aid for Scientific Research from the Ministry of Education, Culture, Sports, Science and Technology, Japan.
- 4Chemotherapy of canine lymphoma with histopathological correlation: Doxorubicine alone compared to COP as first treatment regimen. J Am Anim Hosp Assoc 1987;23:587–596., , , et al.
- 6MOPP chemotherapy for treatment of resistant lymphoma in dogs: A retrospective study of 117 cases (1989–2000). J Vet Intern Med 2002;16:576–580., , , et al.
- 16Quantitative assessment of contaminating tumor cells in autologous peripheral blood stem cells of B-cell non-Hodgkin lymphomas using immunoglobulin heavy chain gene allele-specific oligonucleotide real-time quantitative-polymerase chain reaction. Leuk Res 2003;27:925–934., , , et al.
- 24TNM Classification of Tumours in Domestic Animals. Geneva: World Health Organization; 1980:46–47.