What's new?

Chronic lymphocytic leukemia (CLL) is characterized by both proliferation and accumulation of long‑lived CLL cell clones and various clinical presentation, ranging from a stable to a progressive course. In addition to malignant B cell expansion, a unique feature of CLL is the increased presence of circulating nonmalignant T cells. Here, we found for the first time that CLL progression is associated with increased T‑cell turnover triggered by dysregulated expression of G1 cell cycle phase regulators controlling lymphocyte survival, implying that during CLL development T cells acquire a unique ability to maintain longevity in vivo to be effective in supporting malignant B cell growth. Therefore, our study clearly suggests the involvement of the nonmalignant T‑cell compartment in the disease pathogenesis and clinical outcomes. Understanding the risk factors for CLL development and its clinical consequences is crucial for determining the optimal management of CLL patients.

Introduction

Chronic lymphocytic leukemia (CLL) is a lymphoproliferative disease that manifests as the clonal proliferation of B lymphoid cells with positive expression of CD5, CD19, and CD23 antigens. These cells spread to the peripheral blood (PB), bone marrow, lymph nodes, and extranodal organs, where they accumulate due to their prolonged lifespan.^1,2 With the progression of the disease, lymphadenopathy, hepatosplenomegaly, bone marrow failure, and recurrent infections may occur.^1-3 Patients often develop autoimmune hemolytic anemia or autoimmune thrombocytopenia.^1-3 Despite remarkable phenotypic and cytological homogeneity, the disease exhibits a highly variable clinical presentation and evolution.^1-5 In some patients, the disease progresses slowly and does not require treatment for many years, in contrast to others who experience progression that requires systemic cytostatic treatment soon after a diagnosis is established. Progressive CLL is associated with unique phenotypic and genetic aberrations,^1-3,5-7 and the disease evolution is characterized by a dynamic interplay between the ongoing replication and death within the clones of CLL cells.⁸ CLL patients with higher replication rates are much more likely to develop progressive form of the disease than those with lower replication rates.⁹ Thus, CLL is considered to be a disease of both proliferation and accumulation of long‑lived clones of CLL cells.^8,9

Although the disease affects mature B lymphocytes, patients with CLL usually have abnormalities in the number and function of nonmalignant T cells. Clonal proliferation of malignant B lymphocytes and their accumulation is accompanied by an increase in the absolute number of PB T cells.^10-12 Any apparent T cell lymphocytopenia is due to the diluting effect of the large numbers of CLL cell clones. CLL T cells also exhibit several phenotypic and functional differences, resulting in a defective antileukemia response.^10-13 The correlation observed between the expansion of dysfunctional T cells and PD has led to a suggestion that CLL T cells may create a microenvironment in which the leukemic clones impede differentiation and apoptosis and exhibit increased proliferative activity, sustaining the malignant B cell clones and promoting disease progression.

Since CLL involves both the proliferation and accumulation of long‑lived clones of CLL cells, abnormalities in the expression and function of intracellular factors that regulate proliferation and apoptosis are considered to play a part in CLL pathogenesis. A specific role in the regulation of lymphocyte turnover has been assigned to proteins involved in early G1 phase regulation, cyclin‑dependent kinase inhibitor 1B (p27^Kip1) and cyclin D2.^14,15 Although p27^Kip1 participates in many cellular processes,^16,17 its canonical activity is the inhibition of the majority of known cyclin‑dependent kinases (CDKs), since its crucial role in cell cycle progression.^14,18 Unlike the majority of solid tumors, in which high expression of p27^Kip1 is linked with a favorable prognosis, the increased expression of this protein in CLL cells is associated with a more aggressive clinical course of the disease.^19-21 Alongside p27^Kip1, an increased expression of cyclin D2 was found in CLL cells.^22-24 Cyclin D2 is a member of the D‑type cyclin family involved in the regulation of B lymphocytes passing through the G1 phase restriction point.²⁵ The literature highlights a special role of cyclin D2 in malignant B cells,^22,26 primarily in increasing their longevity in vivo.²³

In contrast to the well‑documented increased expression of p27^Kip1 and cyclin D2 in CLL cells, information concerning their expression in nonmalignant T cells of CLL patients is lacking. The important roles played by p27^Kip1 and cyclin D2 in the regulation of proliferation and apoptosis of CLL cells alongside the massive expansion of nonmalignant T cells prompted us to examine the expression of both G0/G1 phase regulators in relation to ex vivo spontaneous apoptosis of B and T cells in CLL patients. Therefore, our study aimed to compare the apoptosis sensitivity of PB leukemic B and nonmalignant T lymphocytes in relation to the expression of antiapoptotic p27^Kip1 and cyclin D2 proteins in CLL patients with different clinical course as the potential prognostic factors.

Patients and methods

Results

Cyclin‑dependent kinase inhibitor 1B and cyclin D2 expression in CD19+CD5+ and CD3+ cells in chronic lymphocytic leukemia

To determine whether the altered longevity of leukemic CD19+CD5+ and CD3+ cells in CLL might be caused by dysregulation in G0/G1 cell cycle regulator expression, we assessed the p27^Kip1 and cyclin D2 expression in CD19+CD5+ cells and CD3+ cells in the whole CLL group and CLL subgroups and controls (Table 3; Figure 1).

Table 3. Frequency of CD19+CD5+ and CD3+ cells coexpressing cyclin‑dependent kinase inhibitor 1B and cyclin D2 proteins in patients with chronic lymphocytic leukemia and controls

Cell subsetsa	Patients with chronic lymphocytic leukemia			Controls (n = 39)	P value
	Overall (n = 47)	SD group (n = 26)	PD group (n = 17)
Data are presented as median (interquartile range). P values were derived from the nonparametric Kruskal–Wallis 1‑way analysis of variance followed by the Dunn or Mann–Whitney tests. a Percentage of cells with specific phenotype Abbreviations: see Tables 1 and 2
CD19+CD5+p27Kip1+	92.4 (89.3–96.6)	92.2 (89.3–95.9)	93.6 (89.3–97.2)	70.5 (63.8–76.1)	O vs C <⁠0.001 SD vs PD, 0.99 SD vs C <⁠0.001 PD vs C <⁠0.001
CD3+p27Kip1+	95.3 (91.6–96.6)	95.4 (92.4–96.6)	95.6 (91.5–98.2)	73.8 (66.1–79)	O vs C <⁠0.001 SD vs PD, 0.99 SD vs C <⁠0.001 PD vs C <⁠0.001
P value	0.04	0.02	0.07	0.16	–
CD19+CD5+cyclin D2+	46.95 (26.9–96.6)	48.25 (23.6–56)	46.7 (31.4–65.4)	0	O vs C <⁠0.001 SD vs PD, 0.98 SD vs C <⁠0.001 PD vs C <⁠0.001
CD3+cyclin D2+	53.35 (34.8–56.7)	54.65 (51.3–61.8)	49.3 (34.3–67.3)	0	O vs C <⁠0.001 SD vs PD, 0.55 SD vs C <⁠0.001 PD vs C <⁠0.001
P value	0.02	0.004	0.48	–	–

We observed higher proportions of both CD19+CD5+ and CD3+ cells coexpressing p27^Kip1 in all CLL patients, as compared with corresponding healthy cells (P <⁠0.001). In terms of the CLL clinical course, there were no significant differences in p27^Kip1 expression in CD19+CD5+ and CD3+ cells between the SD and PD subgroups. A comparison between p27^Kip1-expressing CD19+CD5+ and CD3+ cells demonstrated that in the whole CLL group, the median frequency of CD3+p27^Kip1+ cells was higher than that of CD19+CD5+p27^Kip1+ cells (P = 0.04), with the difference being more pronounced in the SD group (P = 0.02). In the PD group, there was only a trend toward a higher CD3+p27^Kip1+ cell fraction (P = 0.07). In the controls, no significant differences were observed between the p27^Kip1-expressing lymphocyte subpopulations, and both were significantly lower than the corresponding cells seen in all patient cohorts studied.

We also analyzed the expression of cyclin D2 in the whole group of CLL patients, the clinical subgroups, and the controls (Table 3; Figure 1). Unlike p27^Kip1, which was found in all individuals, cyclin D2 was detected in CD19+CD5+ and CD3+ cells in the CLL patients only. We did not observe any detectable expression of cyclin D2 in healthy CD19+CD5+ nor CD3+ cells. When we compared the patient subgroups regarding the clinical course, we found no significant differences in cyclin D2 expression in CD19+CD5+ and CD3+ cells. A comparative analysis of cyclin D2‑expressing CD19+CD5+ and CD3+ cells showed that in the whole CLL group the fraction of CD3+cyclin D2+ cells was more abundant than that of CD19+CD5+cyclin D2+ cells (P = 0.02) and, of note, this increase was more pronounced in the SD group (P = 0.004). In the PD group, there was no significant difference between CD19+CD5+ and CD3+ cells coexpressing cyclin D2.

Prognostic significance of cyclin‑dependent kinase inhibitor 1B, cyclin D2, and apoptosis in chronic lymphocytic leukemia progression

To assess the prognostic significance of p27^Kip1 and cyclin D2 expression as well as the rate of apoptosis for CLL outcomes, we compared the Kaplan–Meier curves for the disease progression (based on the presence of indications for cytostatic treatment) plotted for 2 patient groups: with the studied variable above (>median, high) and below its median value (≤median, low) (Figure 2).

We noted that CLL progression occurred significantly earlier in the patients with a low proportion of apoptotic CD19+ cells than in those with a high frequency of apoptotic CD19+ cells (HR, 0.2; 95% CI, 0.07–0.57; P = 0.002; Figure 2A). In turn, the fraction of apoptotic CD3+ cells showed no significant association with time to CLL progression (Figure 2B).

Having noted the marked differences in the frequencies of p27^Kip1+ and cyclin D2+ lymphocytes between the CLL patients and the controls, we then examined the association of the regulators’ expression with the clinical course of CLL. We observed, with a borderline level of significance, that the patients with a higher percentage of CD19+CD5+ cells coexpressing p27^Kip1 experienced the disease progression earlier than those with a low fraction of the corresponding cells (P = 0.09; Figure 2C). In addition, as shown in Figure 2F, we noted a tendency toward a shortened time to disease progression in the patients with a low percentage of CD3+ cells coexpressing cyclin D2, as compared with those exhibiting a higher proportion of CD3+cyclin D2+ cells (P = 0.06).

We did not observe any relationship between the expression of p27^Kip1 or cyclin D2 in the CLL cells and patient age or sex, as well as clinical parameters, such as peripheral lymphocyte count, hemoglobin concentration, platelet counts, serum LDH activity, β2‑microglobulin concentration, and Rai / Binet clinical stage.

To clarify the association between p27^Kip1 and cyclin D2 protein expression, the rate of apoptosis, and clinical indices with a risk of CLL progression, we performed the Cox regression analysis (Table 4). In the univariable analysis, among the clinical characteristics studied, including patient age, sex, white blood cell count, lymphocyte count, platelet count, hemoglobin level, and levels of β2‑microglobulin and LDH, we observed that only higher concentrations of β2‑microglobulin predicted CLL progression (P = 0.04), while the presence of mutated immunoglobulin heavy‑chain variable (IgVH) genes was of a protective value (P = 0.002), thus confirming their prognostic significance and enrollment of suitable patients. Among the immune parameters examined, the same analysis allowed us to identify the rate of apoptosis within PB CD19+ and CD3+ cells (P = 0.03 and P = 0.04, respectively) as opposite predictors of the clinical course of CLL. Furthermore, multivariable Cox regression analysis enabled the independent prognostic value for CLL progression to be identified for the IgVH mutated status (HR, 0.16; 95% CI; 0.05–0.52; P = 0.002) as well as apoptosis within both CD19+ (HR, 0.85; 95% CI, 0.72–0.98; P = 0.03) and CD3+ (HR, 1.32; 95% CI, 1.04–1.66; P = 0.02) cells (Table 4).

Table 4. Univariable and multivariable Cox regression analysis for factors influencing time to lymph node and organ progression in chronic lymphocytic leukemia

Prognostic factors	Univariable			Multivariable
	HR	95% CI	P value	HR	95% CI	P value
Abbreviations: see Table 1 and Figures 1 and 2
Apoptotic cells within CD19+ cells, %	0.84	0.72–0.99	0.03	0.85	0.72–0.98	0.03
Apoptotic CD19+ cells within PBMCs, %	0.8	0.63–1.02	0.07	–	–	–
Apoptotic CD3+ cells within PBMCs, %	1.29	1.01–1.65	0.04	1.32	1.04–1.66	0.02
β2‑microglobulin	1.7	1.02–2.81	0.04	–	–	–
IgVH mutation	0.16	0.05–0.52	0.002	0.16	0.05–0.52	0.002

Discussion

Our study demonstrated for the first time that the rate of protection from apoptosis differs between PB B and T lymphocytes in CLL, and that this CLL characteristic may be associated with the clinical course of the disease. An important observation of the study was that there were no significant differences in B cell apoptosis between the patient cohort and healthy controls. Additionally, despite comparable apoptotic B cell fractions in the patients with stable and progressive CLL, the Kaplan–Meyer analysis showed that the lower rate of ex vivo apoptosis of malignant B cells was predictive of earlier CLL treatment implementation, thus indicating the involvement of failed B cell apoptosis in the disease progression. Accordingly, the fraction of apoptotic B cells was also found to be of protective significance when considering the risk of CLL progression. In our ex vivo experiments, we used the CMXRos dye (Molecular Probes) to detect the earliest events in the apoptotic pathway occurring spontaneously within the lymphocytes immediately after cell isolation from PB. In this context, given the small apoptotic compartment within PB B cells, we cannot exclude the long‑lasting influence of prosurvival extrinsic factors from the microenvironment. Our observation is in line with former studies^27-29 showing that leukemic B cells from patients with poor prognosis were more prone to spontaneous apoptosis when the cells were deprived of growth factors in cultures, indicating the presence of stronger prosurvival signals in PB during the disease progression.

Another interesting result of our study was that T cells were found to exhibit the highest rate of spontaneous apoptosis when compared with either leukemic B cells or corresponding healthy T cells. Our results are in line with the study by Bojarska‑Junak et al,²⁷ who demonstrated ex vivo that in CLL the apoptotic fraction was higher in the T cell population than in B cells. In controls, we found that B cells were more prone to apoptosis than T cells, also consistently with previous observations.³⁰ Furthermore, in our ex vivo assay, the rate of T cell apoptosis was shown to be a predictor of earlier CLL progression. This was in opposition to the situation observed within the PB B cell compartment, where the rate of B cell apoptosis demonstrated a protective impact. Accordingly, insignificantly higher fraction of apoptotic T cells was also noted in progressive patients. Taking into account that all samples were incubated under the same conditions, the differences in the apoptosis rate observed between CLL B and T cells strengthen the suggestion that intrinsic factors also play a role in lymphocyte survival, and that resistance to apoptosis is not solely dependent on extrinsic signals. One could speculate that in contrast with B cells, the increased population of PB T cells, as well as ex vivo apoptotic T cells, might indicate that the T cell compartment in CLL becomes more sensitive to intrinsic prosurvival signals than long‑lasting extrinsic ones. Furthermore, the increased T‑cell turnover found in our study may serve to maintain an appropriate size of the T cell population in PB as a relevant support for leukemic B cell growth. This is consistent with recent findings that T cells in CLL patients could prevent apoptosis of malignant B cells.³¹

Our results showing the opposite apoptosis sensitivity of CLL B and T cells prompted us to search for the intracellular factors involved in the maintenance of PB lymphocyte longevity that could be dysregulated in CLL. Therefore, we examined the expression of p27^Kip1 and cyclin D2, which have been shown to be engaged in regulating the G0/G1 phase of the cell cycle by controlling both proliferation and apoptosis pathways.³² We detected a higher population of B and T cells expressing p27^Kip1 inhibitor protein in the CLL patients than in the controls, which demonstrated that a larger compartment of PB lymphocytes in CLL is arrested in the G0/G1 phase irrespective of the clinical course, thus supporting their accumulation in PB. Moreover, the association found between p27^Kip1 expression in leukemic B cells and the earlier progression of CLL seems to emphasize the role of p27^Kip1 in malignant B cell accumulation and CLL clinical course. Similar observations concerning p27^Kip1 expression in leukemic B cells were reported by us and other research teams.^30,31 We did not find any significant differences in leukemic B or T cells expressing p27^Kip1 when considering the CLL clinical course, probably due to the relatively small patient subgroups or methodological differences in the p27^Kip1 assessment. Remarkably, when we compared the p27^Kip1 distribution between B and T cell populations, a novel and surprising observation was made: the highest expression of p27^Kip1 was found in patient T cells, as compared with leukemic B cells and healthy T cells. Available data assessing p27^Kip1 content in CLL are limited to leukemic B cells and / or the overall population of PB lymphocytes,^20,30 and there are no data on p27^Kip1 expression in CLL T cells. In addition, there are several technical differences in the measurement of p27^Kip1 protein expression, meaning that any comparison with our results is difficult. The greater p27^Kip1 abundance in CLL was previously reported to be associated with G0/G1 cell cycle arrest, indicating a quiescent state in PB, increasing with the disease progression and poorer prognosis.^20,30 In contrast, in solid tumors, lower p27^Kip1 abundance in the neoplastic cells was shown to be strongly associated with cell cycle progression, promoting malignant cell proliferation and unfavorable clinical outcomes.^33-36 While the significance of p27^Kip1 in terms of cell cycle arrest and lymphocyte accumulation in CLL is well understood, contradictory results were published regarding the role of p27^Kip1 in apoptotic pathways in cancers showing both an anti- or proapoptotic impact.¹⁶ However, the inverse association of p27^Kip1 with spontaneous apoptosis of PB lymphocytes, as demonstrated in several reports, strengthens the suggestion that in CLL, a high p27^Kip1 expression impedes cell cycle progression, prevents in vivo apoptosis by the inhibition of CDK / cyclin complexes, and compromises patient survival.^20,30 Of note, although we found a higher population of p27^Kip1+ T cells in comparison with p27^Kip1+ leukemic B cells in all CLL patients, the difference was significant only in the SD patients. This finding may confirm that p27^Kip1+ expression in T cells might be involved in the control of CLL clinical outcomes.

Another interesting observation was that cyclin D2 is expressed in CLL patient samples in approximately 50% of the leukemic B and T cells. The expression of cyclin D2 appears to be a unique feature of PB lymphocytes in CLL, as corresponding healthy cells lack cyclin D2 expression.^25,37-39 The lymphocyte expression of cyclin D2 in CLL is considered to be an attribute of cell cycle progression into an early G1 phase, which may, in turn, confer greater sensitivity and responsiveness to the external prosurvival signals.^22,23,40 Recent data reported that all hematopoietic cells died after ablation of D‑type cyclins, demonstrating their antiapoptotic function and necessity for lymphocyte survival.^41,42 Furthermore, a study by Choi et al⁴² showed that the prosurvival action of cyclins D in lymphocytes is displayed through the repression of Fas and FasL expression. Notably, they found that hematopoietic stem cells are protected from apoptosis, and this prosurvival activity is strictly dependent on cyclin D expression even in a quiescent state. This observation is in accordance with current knowledge that cyclins D are induced and function in proliferating cells.⁴³ Therefore, the cyclin D2 expression in CLL found in our study might reflect the cell’s attempt to undergo division upon antigen stimulation. This suggests that cyclin D2+ lymphocytes in CLL exhibit a potency to proliferate and control in vivo apoptosis by inhibition of the extrinsic Fas‑mediated pathway.^42,44 Accordingly, recent studies have demonstrated that in CLL, lymphocytes from proliferating compartments show a greater ability to undergo spontaneous apoptosis in vitro.^27,28 Likewise, we and other authors have previously reported that the expression of cyclin D2 in vivo might be a marker of leukemic B cell activation.^22,23,43 Moreover, we had showed earlier that the rate of decrease in cyclin D2 expression observed in cell cultures corresponded with an increase in spontaneous apoptosis of leukemic B cells, thus clearly indicating that cyclin D2 might indeed prevent apoptosis and support survival of PB leukemic B cells.

Remarkably, while we observed higher cyclin D2 expression in a subset of T cells, as compared with leukemic B cells, the difference was significant only in patients with SD, similarly to the findings regarding p27^Kip1 expression. In line with this observation, we also found that greater abundance of cyclin D2 within T cells may predict a stable CLL course. Taking into account that p27^Kip1 and cyclin D2 have been considered as proteins with an antiapoptotic function in CLL, our results may indicate that PB T cells from stable CLL patients are more protected from apoptosis than those from patients with PD, thus implying an association between the longevity of T cells and CLL outcome. Therefore, our study demonstrates for the first time that despite p27^Kip1 and cyclin D2 increase in both CLL patient cohorts, the loss of differences in their expression between the leukemic B and T cells during CLL progression may lead to the induction of the apoptotic pathway in the larger T cell compartment. Our observation appears to be consistent with studies in CLL, indicating that the lymphocytes with higher proliferative potential are more prone to ex vivo apoptosis.²⁷ Based on these findings, we suggest that in CLL, the dynamic interplay of p27^Kip1+ and cyclin D2+ fractions within B and T cells may play a role in their diverse resistance to apoptosis and, in consequence, determine the clinical course of the disease.

The main limitation of our study is the relatively small size of both patient and control groups. Therefore, it is possible that some relationships present in the general population were not detected with the statistical tests used. Moreover, it must be noted that the CLL cells that undergo apoptosis in vivo are quickly eliminated from the circulation. For this reason, the ex vivo measure of apoptosis in CLL lymphocytes, whatever the method, probably does not accurately reflect what happens in vivo. Thirdly, the impact of the studied phenomena on the clinical course of CLL could be assessed only regarding the rapidity of the disease progression before treatment. Since the completion of patient recruitment into the study, the standards for anticancer treatment of CLL have substantially changed (immunochemotherapy is rarely used), so the assessment of the effects the parameters tested in our study had on the overall survival would be of very limited practical value.

In conclusion, our study showing the high expression of p27^Kip1 and cyclin D2 cell cycle regulators in quiescent lymphocytes in CLL patients strengthens the suggestion of the induction and maintenance of prosurvival signals in leukemic B and nonmalignant T cells. This may be one of the mechanisms underlying the accumulation of both lymphocyte subpopulations in PB of the CLL patients. Taking into account that the fractions of p27^Kip1+, cyclin D2+, and ex vivo apoptotic cells were higher in T cells than B cells, we may speculate that T cells during CLL development acquire the unique ability to maintain longevity in vivo to be effective in supporting malignant B cell growth. Our finding on the predictive significance of the spontaneous apoptosis of T cells for CLL progression indicates the involvement of the nonmalignant T cell compartment in CLL pathogenesis and clinical outcomes.