Recurrences of Superficial Bladder Carcinoma Are Associated with a Raise of Cd8 High Cd57 + and Cd8 Low T Lymphocytes in Peripheral Blood

Immunotherapy with BCG is effective in patients with recurrent superficial bladder carcinoma. This therapy involves Interleukin-2 (IL-2), but little is known about the immunological parameters involved in superficial bladder car-cinoma. We have monitored immunological parameters in twenty patients with superficial bladder carcinoma treated with transurethral resection (TUR) followed by IL-2 instillation. Cell numbers of peripheral blood leukocyte subpopulations were counted before surgery and during follow-up after surgery. During follow-up, we compared the cell counts in patients with and without a recurrent tumour. We used the values of healthy matched controls as a reference. Recurrent disease in patients corresponded with a significant increase in CD8 + lymphocytes, and especially the CD8 high CD57 + and CD8 low subpopulations. The phenotype of these T lymphocytes belongs to cells with an immunosuppressive function. We hypothesize that these peripheral immune suppressive cells facilitate tumour recurrences or that tumour recurrences cause an increase in peripheral immune suppressive lymphocytes.


INTRODUCTION
Local immunotherapy is effective in delaying recurrences of superficial bladder carcinoma after tumour removal by transurethral resection (TUR).In patients with high-risk superficial bladder carcinoma, immunotherapy with Bacillus Calmette-Guérin (BCG) is an important therapeutic intervention superior to chemotherapy [1][2][3].This underlines the role of the immune system in control of local disease.However, little is known of the underlying immunological mechanisms.
In this paper we explore cell numbers of peripheral blood leukocyte populations in patients with bladder carcinoma.These patients participated in a phase II study of the efficacy of intravesical instillations of IL-2 after complete TUR.To identify indications of changes in innate or antigen-specific leukocytes related to recurrent bladder carcinoma, we compared data from patients at the time of TUR with data of agematched controls, and patients at later time-points with and without tumour recurrences.

Patients and Controls
The study protocol was approved by the Lithuanian Bioethics Committee and State Medicines Control Agency of Lithuania.Patients with histologically confirmed primary Ta/T1 non-muscle invasive bladder carcinoma were included into the study.All study patients gave written informed consent.Criteria for exclusion were (a) white blood cells < 3,000/mm 3 or platelets < 100,000/mm 3 ; (b) hepatic enzymes (SGOT, SGPT, alkaline phosphatase) or creatinine > 2x normal values; (c) previous chemotherapy or radiotherapy within 3 months before treatment; (d) previous or concurrent cancer at other sites; (e) patients with urinary tract infection; (f) patients with tumours located in the prostatic urethra or in a diverticulum.
The control group consisted of age-and sex-matched healthy subjects.These subjects had no history of any oncological, autoimmune diseases, chronic infections (HIV or HCMV) or alcoholism.

Treatment and Follow-Up
Patients were treated with instillations containing 9x10 6 IU recombinant IL-2 (Chiron, Amsterdam, the Netherlands; nowadays Novartis) on 5 consecutive days, beginning on the second day after TUR.The IL-2 was diluted in 50 ml saline (0.9% NaCl) and instilled in the bladder through a catheter, which was removed after the instillation.The drug remained in the bladder for 1-2 hours.The first follow-up cystoscopy was performed about 2 months after TUR.Further cystoscopies were performed at periodic visits to the urologists, all suspicious lesions were resected and subjected to histological examination.

FACS Analysis of Peripheral Blood Lymphocyte Subsets
Blood samples were drawn for flow cytometric analysis prior TUR (visit 0) and during two follow-up visits that were on average 2.0 (range 1.2 to 3.4) and 8.3 (range 3.2 to 16.3) months after TUR.The samples of peripheral blood were analyzed on a FACSort (Becton Dickinson) flow cytometer with a laser tuned at 488 nm.The lymphocytes were stained with CD3-FITC/CD16/56-PE and CD57-FITC/CD8-PE/CD4-PerCP combinations of fluorochrome conjugated monoclonal antibodies (Becton Dickinson).Data were acquired and analyzed with CellQuest software (Becton Dickinson).Forward and side scatter were used to gate the lymphocytes.List mode files were collected for 10 4 cells from each sample.Percentages of CD3 -CD16/56 + , CD3 + CD16/56 -, CD4 + and CD8 + lymphocyte subpopulations were determined using conventional flow cytometric analysis.The subsets of CD8+ lymphocytes (CD8 high CD57 + , CD8 high CD57 -and CD8 low ) were defined as described in our previous publication [35].

Statistics
Data of patients were grouped according to the presence or absence of a tumour recurrence at sampling.Statistical significance was analyzed by Students t-test for normally distributed data.If data differed by more that 2-fold between groups, then data were logarithmically transformed prior to analysis.This procedure did not affect analyses crossing the limit of statistical significance.Statistical analysis was performed progressively starting at general leukocyte populations, and focussing on the subpopulations of the cell types that showed significant differences between both groups.The lower limit of statistical significance was set at p<0.05 (two-sided interval).

General Characteristics
Between May 2005 and June 2006, we enrolled twenty consecutive patients with histologically confirmed Ta nonmuscle invasive bladder carcinoma.Only patients were included from which immunologic parameters were assessed prior surgery.There were 4 females and 16 males with the average age of 62 years (range 43 to 82 years).These patients were compared with 22 matched controls, 5 females and 17 males with an average of 62 years (range 43 to 84 years).
The selected patients had primary tumours including 5 with TaG1, 12 TaG2, 3 TaG3.Statistical analyses showed no significant differences in peripheral leukocyte subpopulations in patients with low or high grade tumours (data not shown).

Recurrence-Free Versus Recurrent Disease
We compared leukocyte subpopulations in 13 samples taken in patients without and 13 samples taken from patients with recurrences.Six patients had no recurrence at the first visit, but had a recurrence at the second visit; so they contributed to both groups.differ in the various groups.So we further analysed the lymphocyte subpopulations.
Fig. (2) shows that patients with recurrences have higher CD8 + lymphocyte counts than matched healthy controls, patients at TUR, and recurrence-free patients.Especially the increases in patients with recurrences compared to matched controls (76%) and recurrence-free patients (69%) are considerable.Numbers of CD16/CD56 + and CD4 + did not differ.So we further analysed the CD8 + subpopulations.Fig. (3) shows that the CD8 + increases reside in CD8 high CD57 + and CD8 low subpopulations.Patients with recurrent disease have increases of 217% and 180% in CD8 high CD57 + lymphocyte numbers compared to matched healthy controls and recurrence-free patients, respectively.In patients with recurrences the numbers of CD8 low were 71% increased compared to matched controls.

Immunology before or after Recurrence
Tumours and the immune system have bidirectional interaction.Firstly, tumours could induce immune suppression, and secondly immune suppression could facilitate tumour growth.If tumour growth would cause immune suppression, we expect immunological changes when the tumours become detectable, compared to earlier measurements.This can be tested by comparing lymphocyte numbers before and at tumour recurrence in the same patients.If immune suppression causes tumour recurrence, we expect immune suppression to be present, prior tumour recurrence.This can be tested by comparing the data at surgery for patients that have an early or late tumour recurrence, i.e. a recurrence before or after nine months.significant increases were found of 51% and 71% in CD8 + and CD8 low lymphocytes, respectively.The CD8 high CD57 + were increased by 57%, albeit not significant.The increase of suppressor cells during tumour recurrence suggests that the tumour induces immune suppressor cells.5) compares immunological differences present at surgery to see if these foreshadow tumour recurrence.This question was assessed by analyzing twelve patients without a recurrence at nine months, and eight with a recurrence at nine months.Patients with a recurrence within nine months had higher values of CD8 + , CD8 high CD57 + , CD8 low , of 32%, 56%, and 57%, respectively, albeit not statistical significant.The higher number of suppressor cells at surgery is compati-ble with the concept that immune suppression facilitates tumour growth.Fig. (6) summarizes the data of the previous two Figures.Increases of 50 to 75% were found in the CD8 high CD57 + and CD8 low subpopulations in both analyses.This overview illustrates that the increases in CD8 + lymphocyte subpopulations partly precede tumour growth, but that tumour growth precedes additional increases in CD8 + lymphocyte subpopulations.

DISCUSSION
Our study highlights numerical changes in leukocyte populations during recurrence of superficial bladder carcinoma.Peripheral blood leukocyte populations that were in- ).Differences for the CD8 high subpopulations, were not significant, i.e. p=0.12 and 0.17, for CD57 + and CD57 -cells, respectively.

Fig. (5). Differences between CD8
+ lymphocyte subsets at baseline in patients with or without a recurrence at 9 months.Data shown of patients at surgery.The patients are grouped according to the presence of a recurrence within nine months (horizontal lines; n=12) or absence of a recurrence at nine months (vertical lines; n=8).No values of p<0.05 were found, but p= 0.21, 0.20, 0.78, 0.33, for all CD8 + , CD8 high CD57 + , CD8 high CD57 -and CD8 low cells, respectively.Absolute counts per ul creased in patients with recurrences were CD8 + lymphocytes, most notably the CD8 high CD57 + and CD8 low subsets.
Different lymphocyte subsets are differentially distributed in peripheral blood, (tumour) tissue and the lymph compartment [36].Thus, changes in cell numbers in peripheral blood may not reflect changes in total cell numbers per patient.In patients with malignant tumours, increased numbers of CD57+ cells were found in the draining lymph node [37].If this is also the case in our patients, than the increase found would actually be an underestimate of the total increase in bladder cancer patients.
The therapeutic effects of intravesical application of IL-2 after complete tumour removal appear to be minimal in this trial (manuscript in preparation).This contrasts with earlier studies [20][21][22][23][24][25][26][27] that show therapeutic effects after incomplete TUR.IL-2 is most effective, if IL-2 interacts directly with the tumour [13].This interaction would be possible after incomplete TUR, but not in the present study, because the tumours were completely removed.Although in this study, IL-2 lacked clinical efficacy, it is not clear if it affected immunological parameters.
Two other groups have also shown the rise in peripheral CD8 + lymphocytes in patients with bladder carcinoma compared to healthy controls [38,39].However, a third study by Agarwal and co-workers obtained different results.They reported a decrease in percentage CD3 + , CD4 + , CD8 + and CD56 + cells in patients compared to healthy controls [40].Two remarkable differences between this paper and the other papers should be noticed.Firstly, Agarwal did not use agematched controls.However, total leukocytes, neutrophils and NK cells increase with age [41,42], which results in a decrease in percentage of other leukocyte populations.Secondly, instead of using a standard region, as is done in this study and by most other groups do [38,39], Agarwal and coworkers drew manually a lymphocyte region for each analysis [40], which is less reproducible.CD8 + lymphocytes are best known as cytotoxic T lymphocytes.However, a low expression of CD8 is a feature of anergic T lymphocytes [43] that have low mRNA levels of molecules involved in cytotoxicity and proinflammatory cytokines [33,44].In transplantation biology, these cells are known as suppressor T lymphocytes [34].CD57 expression on CD8 + lymphocytes is a general marker of proliferative inability [45] and susceptibility to apoptosis [46].These cells are also associated with immunosuppression [47,48].In cancer patients, CD8 high CD57 + subsets suppress cytotoxic T lymphocyte functions [29,30,32].In brief, both CD8 low and CD8 high CD57 + subsets are associated with antigen-specific suppressor T cell function.CD8 + CD57 + suppressor cells are generally CD28 - [49,50], and in line with results of this paper, CD8 + CD28 -lymphocytes have been found to decrease after removal of bladder carcinoma [51].
Previously, we have shown that increased numbers of peripheral CD8 high CD57 + lymphocytes are associated with negative prognosis for survival in melanoma patients treated with interferon-therapy [52].Renal cell carcinoma patients with higher numbers of CD8 high CD57 + (suppressor) cells had a shorter survival.However, renal cell carcinoma patients with high CD8 high CD57 + numbers responded better to interferon-therapy [35].
Our data show a correlation between T lymphocyte subpopulations in peripheral blood and recurrence of bladder carcinoma.The involved T lymphocytes (CD8 low and CD8 high CD57 + ) have molecular markers of suppressor T lymphocytes.Although technically challenging, it will be interesting to test if these cells are also functionally suppressor T lymphocytes.Another challenge would be to show that the increase of the suppressor T lymphocytes is not just a redistribution of cells to the peripheral blood, but also occurs in the tumour.
Considering the role of suppressor cells, it would also be interesting to monitor numerical changes in CD4 + Foxp3 + CD25 + T suppressor cells [53,54].These suppressors are only a small fraction (approximately 5%) of CD4 + T cells.Thus, our progressive analysis (Fig. 2) might have overlooked numerical changes in a small fraction of CD4 + cells.(4 and 5).Percentage increase in cell numbers after surgery in patients that have a recurrence at 9 months (squares) versus those that do not have a recurrence at 9 months (diamonds).

CONCLUSION
Tumour growth and tumour recurrence are accompanied by increases of CD8 high CD57 + and CD8 low lymphocytes.It remains unclear if these cells precede or follow tumour recurrence, most likely it is a combination of both.These cells are probably immune suppressor cells.

Fig. ( 1 Fig. ( 1 )
Fig.(1) shows slightly increased lymphocyte counts in patients with recurrent disease compared to matched healthy controls.The numbers of monocytes and neutrophils did not

Fig. (
Fig. (5) compares immunological differences present at surgery to see if these foreshadow tumour recurrence.This question was assessed by analyzing twelve patients without a recurrence at nine months, and eight with a recurrence at nine months.Patients with a recurrence within nine months had higher values of CD8 + , CD8 high CD57 + , CD8 low , of 32%, 56%, and 57%, respectively, albeit not statistical significant.The higher number of suppressor cells at surgery is compati-

Fig. ( 4 )
Fig. (4).Changes in CD8+ lymphocyte subset counts within patients at surgery and recurrence.Data of CD8 + lymphocyte subsets of the same ten patients at TUR (grey) and at tumor recurrence (black) are shown.Values of p<0.05 are indicated by a (p=0.045),b (p=0.018).Differences for the CD8 high subpopulations, were not significant, i.e. p=0.12 and 0.17, for CD57 + and CD57 -cells, respectively.

Fig. ( 6 )
Fig.(6).Relative increase in CD8 + subsets before and after recurrence.Graphical summary of data shown in Figs.(4 and 5).Percentage increase in cell numbers after surgery in patients that have a recurrence at 9 months (squares) versus those that do not have a recurrence at 9 months (diamonds).