Decreased Absolute Counts of T Lymphocyte Subsets and Their Relation to Disease in Squamous Cell Carcinoma of the Head and Neck

Vol. 10, 3755–3762, June 1, 2004 Clinical Cancer Research 3755

Iris Kuss,1 Bridget Hathaway,3 Conclusions: Patients with SCCHN have altered lym- Robert L. Ferris,1,3 William Gooding,1 and phocyte homeostasis, which persists for months or years after curative therapies. Theresa L. Whiteside1,2,3 1 University of Pittsburgh Cancer Institute and Departments of INTRODUCTION 2Pathology and 3Otolaryngology, University of Pittsburgh School of Medicine, Pittsburgh, Pennsylvania Despite the constant input of new lymphocytes from the bone marrow and the thymus and the exponential generation of specific cells in response to antigens, the size of the peripheral ABSTRACT lymphocyte pool remains relatively stable. To maintain home- Purpose: Apoptosis of circulating CD8؉ T cells seen in ostasis, the peripheral blood T-lymphocyte pool is regulated in patients with squamous cell carcinoma of the head and neck complex ways that have not yet been fully defined (1). In [SCCHN (Hoffmann T, et al. Clin Cancer Res 2002;8:2553– general, clonal expansions of specific cells are balanced by 62)] suggested a possibility of lymphocyte imbalance. There- programmed cell death in other subpopulations and the mainte- fore, absolute numbers and percentages of lymphocyte sub- nance of a relatively constant total peripheral blood T-lympho- sets were examined in the peripheral blood of SCCHN cyte count. We have reported previously that patients with patients and controls. squamous cell carcinoma of the head and neck (SCCHN) as well Experimental Design: Venous blood was obtained from as melanoma and breast cancer have increased proportions of 146 patients with SCCHN and 54 normal volunteers. Abso- circulating T cells that bind annexin V and are, therefore, lute numbers of CD3؉, CD4؉, and CD8؉ T lymphocytes destined to apoptose (2–5). Due to this extensive apoptosis in were determined using fluorobeads in a flow cytometry- the peripheral circulation and concomitant repopulation of the based technique. Percentages of T lymphocyte subsets were blood compartment with T cells from the immature cell pool, also evaluated by flow cytometry. The patients were grouped rapid turnover of effector cells takes place that is reminiscent of at the time of blood draw [active versus no evidence of that described for patients with HIV (6, 7). These observations disease (NED), type of therapy administered, and the length suggest that the evaluation of the proportions and absolute numbers of lymphocytes and their subsets in the peripheral of follow-up]. circulation of cancer patients is important and might provide Results: Patients with SCCHN had significantly lower insights into redistribution of those lymphocyte subsets that absolute numbers of CD3؉ CD4؉, and CD8؉ T cells than mediate antitumor defense. normal controls. However, no differences in the percentages Patients with cancer who do not receive conventional ther- of T-cell subsets between patients and normal controls were apies are not generally considered to be lymphopenic. Never- observed. Patients with active disease had significantly lower theless, abnormalities in T-cell counts have been observed in -CD3؉ and CD4؉ T-cell counts than those with NED. Pa patients with breast cancer, ovarian cancer, myeloma, head and tients who had NED after surgery and radiotherapy had the neck cancer, or liver cancer (8–11). More importantly, some lowest T-cell counts among the NED cohort. Patients who reports show associations of T-cell subset abnormalities with had NED for >2 years did not recover their T-cell counts, poor clinical outcomes (11–13). However, there is no consensus and the T-cell imbalance was evident many years after on the extent of changes in different T-lymphocyte subsets in curative surgery. The tumor-node-metastasis (TNM) stage the course of cancer progression or its relationship with re- or site of the disease was not related to the absolute T-cell sponse to tumor-specific therapy or patient survival. count. Patients with recurrent disease at the time of blood Absolute numbers of lymphocyte subsets in the peripheral ؉ draw tended to have the lowest CD4 T-cell counts. circulation have been traditionally measured by dual-platform technologies, which couple percentages of positive cell subsets determined by flow cytometry with the absolute lymphocyte count obtained by automated hematology analyzers. Until recently, this was a standard, universally used technology, and it Received 1/9/04; revised 2/18/04; accepted 2/23/04. The costs of publication of this article were defrayed in part by the has been suggested that it is responsible for substantial differ- payment of page charges. This article must therefore be hereby marked ences in absolute lymphocyte counts reported by different lab- advertisement in accordance with 18 U.S.C. Section 1734 solely to oratories (14, 15). The more recent development of single- indicate this fact. platform technologies, which are performed entirely on the flow Requests for reprints: Theresa L. Whiteside, University of Pittsburgh Cancer Institute, Research Pavilion at the Hillman Cancer Center, Suite cytometer, has significantly improved the assay precision and 1.27, 5117 Centre Avenue, Pittsburgh, PA 15213, Phone: (412) 624- accuracy and allowed for greater uniformity of results between 0096; Fax: (412) 624-0264; E-mail: [email protected]. laboratories (14–16).

To evaluate a possible impact of spontaneous apoptosis of Table 1 Clinicopathological characteristics of the patients with circulating T lymphocytes (2–5) on the peripheral T-cell pool, SCCHNa included in the study we investigated both the percentages and absolute numbers of Characteristics N ϩ ϩ ϩ CD3 , CD4 , and CD8 T lymphocyte subsets in a large Total patients 146 cohort of patients with SCCHN, using a single-platform flow Age (yrs) cytometry-based method. We found that the presence of tumor Mean (range) 62 (24–86) b as well as its recurrence had a significant impact on the absolute Tumor site Oral cavity 58 number of T-cell subsets. Both the type of therapy and length of Larynx 38 the posttherapy period were considered, and the results sug- Pharynx 29 gested that regardless of these factors, lymphopenia was a Hypopharynx 10 persistent feature of the disease. Other/unknown 11 Tumor stageb

T1 55 T 27 MATERIALS AND METHODS 2 T3 14 Patients and Controls. In total, 146 patients with SC- T4 26 CHN, who were seen consecutively between July 2001 and June CIS 1 2003 at the Outpatient Otolaryngology Clinic at the University Unknown 23 Nodal statusc of Pittsburgh Oral Cancer Center, were entered into the study. 072 The Institutional Review Board has approved the protocol for 124 collection of patient blood samples. Normal healthy donors 228 [normal controls (NCs)] were recruited among the laboratory 32 Tumor differentiationc personnel, family members of patients, and other volunteers, Well 19 with an intent to match controls and patients for age. Subjects Moderate 95 who served as NCs were interrogated for the general state of Poor 21 health, use of medications, smoking, and alcohol consumption. Undifferentiated 1 Written informed consent was obtained form each individual CIS 1 Unknown 9 participating in this study. Active disease 48 The characteristics of all of the patients included in this Before surgery for primary tumor 30 study are shown in Table 1. The cohort of 146 patients included Before surgery for recurrent tumor 10 nd 109 men and 37 women with a median age of 62 years (range, Before surgery for 2 primary tumor 2 Recurrent disease 6 24–86 years), and the group of 54 volunteers comprised 25 No active disease 98 males and 29 females with a median age of 54 years (range, After surgery alone 56 22–88 years). Of 146 patients, 48 had active disease, and 98 had Ͻ2 yrs 26 no evidence of disease (NED) after surgery alone (n ϭ 56) or Ͼ2 yrs 30 after surgery and radiotherapy (n ϭ 42; Table 1). After surgery and RT 42 Ͻ2 yrs 25 Collection of Blood Samples. Venous blood (10 ml) was Ͼ2 yrs 17 obtained from the patients either the morning before surgery or Smoking history during the outpatient visits and before any therapies. Blood Yes 92 samples were collected into heparinized tubes. Samples were Former 16 No 28 also collected from age-matched NCs. Blood samples were Unknown 10 hand-carried to the laboratory and immediately used for exper- a SCCHN, squamous cell carcinoma of the head and neck; CIS, iments. carcinoma in situ; RT, radiation therapy. Absolute Lymphocyte Count Determination. A stand- b At the time of initial diagnosis. ard single-platform technique, tetraONE System (Beckman c Determined at the time of surgery. Coulter, Miami, FL), based on four-color flow cytometry in the presence of counting beads was used. The identification of lymphocytes by expression of bright CD45 and low side scatter signals was followed by the identification of T-cell subsets along with cells. The sample acquisition and analysis were based on the expression of CD3, CD4, and CD8, as described performed on the EPICS CL flow cytometer with a fully auto- previously (16). Briefly, 100 ␮l of heparinized anticoagulated mated software-reagent combination. The number of cells (or blood were incubated with 10 ␮l of tetraCHROME reagent cell subsets) per microliter was obtained by calculating the containing anti-CD45-FITC-, anti-CD4-RD1-, anti-CD8-ECD-, number of cells counted ϫ concentration of beads/number of and anti-CD3-PE-Cy5 (Beckman Coulter)-labeled antibodies. beads counted. In each experiment, blood obtained from patients Specimens were then lysed with the ImmunoPrep Reagent Sys- was evaluated together with blood from at least one NC. The tem at the Coulter TQ-Prep Workstation. Leukocyte morphol- one-platform method was validated in our laboratory by com- ogy and cell surface integrity were maintained by a gentle, parisons with the previously established two-platform method. no-wash erythrocyte lysing method. Immediately before analy- Statistical Analysis. Differences in percentages and sis, 100 ␮l of Flow-count Fluorospheres (Beckman Coulter) counts of lymphocyte subsets between patients and NCs were were added to the stained cells, and the beads were counted age-adjusted by linear regression models after suitable data

Fig. 1 Percentages of CD3ϩ (A), CD4ϩ (B), and CD8ϩ (C) T lymphocytes in the peripheral circulation of patients with squamous cell carcinoma of the head and neck and normal controls distributed according to age. The horizontal axes show a rug plot of case incidence by age (1 tick/subject). The top axis displays normal controls, and the bottom axis displays patients.

transformations. If age adjustment was unnecessary, differences the absolute counts of CD3ϩ and CD4ϩ but not CD8ϩ T cells were tested with the t test or the Wilcoxon test. The associations decreased significantly with age in patients and NCs. As shown among lymphocyte subsets and clinical (disease status and site in Fig. 2, the absolute counts for CD3ϩ cells decreased for of disease), pathological (stage), and behavioral (smoking) char- patients and NCs with an estimated slope of Ϫ7.3 (P ϭ 0.02) acteristics were tested with the t test or Wilcoxon test for between the ages of 24 and 88 years. CD4ϩ cells decreased with two-group differences or the Kruskal-Wallis test for three or a slope of Ϫ5.5 (P ϭ 0.01). CD8ϩ T cells were unrelated to age more groups. Tests of trend with ordinally scaled end points in either group (P ϭ 0.66). such as T and N stages were conducted with Jonckheere-Terp- In comparing patients with NCs, no significant differences stra test. were observed in the percentages of T-cell subsets (Fig. 1; Table 2). Age-dependent comparisons of absolute lymphocyte counts RESULTS in patients versus NCs were adjusted for age. The patients had Analysis of T-Cell Subsets in Patients and NCs. Using 382 fewer CD3ϩ cells, 297 fewer CD4ϩ cells, and 85 fewer the single-platform method, we initially compared both the CD8ϩ cells per mm3 than NCs. These significant differences percentages and absolute numbers of CD3ϩ, CD4ϩ, and CD8ϩ (P Ͻ 0.0001 for CD3ϩ and CD4ϩ cells; P Ͻ 0.03 for CD8ϩ cells in all patients with those in NCs. Because NCs were, on cells) were detectable at any age (Fig. 2). Overall, the patients average, 8 years younger than patients, we first examined had significantly lower absolute T-cell counts but not percent- whether the percentages and counts of T lymphocytes changed ages than NCs, and the absolute numbers of CD8ϩ T cells were with age to be able to correctly control for age differences when the least depressed in patients with SCCHN relative to NCs. comparing patients with NCs. We found that the percentages of The CD4/CD8 Ratio in Patients and Controls. The T lymphocytes were not associated with age (Fig. 1). However, mean CD4/CD8 ratio was not significantly different between

Fig. 2 Absolute counts of CD3ϩ (A), CD4ϩ (B), and CD8ϩ (C) T lymphocytes in the peripheral circulation of patients with squamous cell carcinoma of the head and neck and normal controls distributed according to age. As indicated by the regression lines, numbers of CD3ϩ and CD4ϩ T cells decreased with age in normal controls and patients with squamous cell carcinoma of the head and neck (P ϭ 0.02 and 0.01, respectively). However, the patients had significantly lower numbers of CD3ϩ, CD4ϩ,orCD8ϩ T cells than normal controls at any age (P Ͻ 0.0001 for CD3ϩ and CD4ϩ cells; P Ͻ 0.03 for CD8ϩ cells). The counts of CD8ϩ T cells were not related to age, and the horizontal lines in C depict the means for the number of CD8ϩ T cells in each cohort.

Table 2 Absolute numbers versus percentages of T-cell subsets in the A somewhat different relationship emerges when CD4ϩ peripheral circulation of patients with SCCHNa and NCs and CD8ϩ T cells are examined in the patients who, at the time Data are means Ϯ SD. of blood draws, had NED either after curative surgery or after CD3ϩ CD4ϩ CD8ϩ surgery plus radiotherapy (Fig. 3C). Here, it can be seen that Absolute no. patients previously treated with radiation therapy (RT) in addi- Patients (N ϭ 146) 966 Ϯ 605 673 Ϯ 415 391 Ϯ 270 tion to surgery had the lowest CD4ϩ counts. Other patients with NCs (N ϭ 54) 1388 Ϯ 494 1005 Ϯ 360 476 Ϯ 208 NED had somewhat depressed CD4ϩ counts but normal CD8ϩ Ͻ b Ͻ b P 0.0001 0.0001 0.0343 counts when compared with NCs (Fig. 3A). Percentages Patients (N ϭ 146) 71 Ϯ 11 44 Ϯ 11 26 Ϯ 11 Effects of Postoperative Radiation or Chemotherapy on NCs (N ϭ 54) 70 Ϯ 947Ϯ 922Ϯ 7 Absolute Lymphocyte Counts. Among 98 patients with P 0.7158 0.1462 0.0404 NED at the time of blood draws, 42 had received postoperative CD4/CD8 ratios RT. The postoperative RT consisted of 66–70 Gy divided into Patients (N ϭ 56) 2.1 (1.2) NCs (N ϭ 54) 2.4 (1.1) 1.8–2.0-Gy fractions over 35 individual daily fractions. This RT P 0.2012 was applied within 6 weeks of surgery. No patient received a SCCHN, squamous cell carcinoma of the head and neck; NC, hyperfractionated RT. The time duration between last RT and a normal control. blood draw for this study was variable, ranging from 1 to 302 b Test of differences between patients and controls is adjusted for months (mean Ϯ SD, 47 Ϯ 67 months). The patients who had age. RT also had significantly lower absolute numbers of CD4ϩ T cells (P Ͻ 0.0001) as well as CD3ϩ T cells (P ϭ 0.0043) compared with 56 patients with NED without prior RT (Fig. 4). patients and NCs (P ϭ 0.21; Table 2). However, the same In contrast, the count of CD8ϩ T lymphocytes remained un- CD4/CD8 ratio does not discriminate between patients with a changed after RT, an indication that CD8ϩ T cells were not high versus low absolute count of the two lymphocyte subsets. sensitive to RT. Nevertheless, it should be noted that 37 of 56 In our series, for example, one patient with a CD4/CD8 ratio of patients with NED treated with surgery alone also had compa- 2.44 had an absolute count of 800 CD4ϩ cells and 328 CD8ϩ rably low CD4ϩ and CD8ϩ T-cell numbers, and, thus, RT was cells, whereas another patient with the same CD4/CD8 ratio of only one of the factors contributing to T-cell cytopenia in the 2.44 had 155 CD4ϩ cells and 64 CD8ϩ cells. For this reason, patients. we undertook a more detailed analysis of the relationship be- Sixteen of 42 patients treated with surgery and radiother- tween the numbers of CD4ϩ and CD8ϩ T lymphocytes in apy also received chemotherapy, which was concluded at least individual patients, as shown in Fig. 3. Relative to NCs, whose 2 months before the blood draws for this study. The mean time mean CD4/CD8 ratio is 2.4 (Fig. 3A), patients with active period elapsed between last chemotherapy and the flow analysis disease, whose mean CD4/CD8 ratio is 2.3, have decreased was 18 Ϯ 27 months (range, 2–85 months). There was no numbers of CD4ϩ and CD8ϩ T cells. As a result, there is a difference in the CD4ϩ T cell (P ϭ 0.36) or CD8ϩ T cell (P ϭ dramatic shift of the individual CD4/CD8 ratios to the bottom 0.61) counts between patients receiving chemotherapy plus ra- left quadrant in Fig. 3B. Not surprisingly, the patients with diotherapy and patients treated with radiotherapy alone. recurrent disease have the lowest absolute counts of CD4ϩ and Effects of Surgery on Absolute Lymphocyte Counts. CD8ϩ T cells (Fig. 3B). Among the patients with active disease, The cohort of 56 patients with NED who underwent curative a small subgroup (8 of 48 patients) with high numbers of CD4ϩ surgery alone was divided into those treated Ͼ2 years before the T cells and relatively normal counts of CD8ϩ T cells (top right blood draw and those studied within 2 years of surgery. As seen quadrant in Fig. 3B) is evident. in Fig. 5, the absolute numbers of the T-cell subsets were

Fig. 3 Distribution of the individual CD4ϩ and CD8ϩ T-cell numbers among normal controls (A), squamous cell carcinoma of the head and neck patients with active disease (B), and squamous cell carcinoma of the head and neck patients with no evidence of disease (NED; C). The horizontal and vertical reference lines indicate mean values for CD4ϩ and CD8ϩ T cells of normal controls tested in the same assays as the patients. The circles indicate prevalent distribution of cell counts. Note the very low numbers of both CD4ϩ and CD8ϩ T cells in B and C.InB, patients treated surgically for disease recurrence or second primary tumors are indicated by ϩ.InC, patients treated with postoperative radiotherapy are denoted by a triangle.

Fig. 4 Box plots showing effects of postoperative radiation therapy (RT) on the absolute numbers of CD3ϩ, CD4ϩ, and CD8ϩ T cells in patients with squamous cell carcinoma of the head and neck. The absolute counts of patients who received RT (n ϭ 42) are compared with counts in the patients treated with surgery without RT (n ϭ 56). The bars are median values, the box indicates the interquartile range (25–75%), and the whiskers extend to 1.5ϫ the interquartile range.

significantly decreased in both groups relative to NCs (see Fig. surgery, whereas that of CD8ϩ T cells recovered, coming up to 3A; P Ͻ 0.0001). The data suggest that effects of the tumor on the normal range in all patients except those with stage IV the homeostasis of lymphocytes are observed long after the disease (data not shown). tumor is removed and in the absence of any other lymphoabla- Lymphocyte Counts and Disease. Among 42 patients tive therapy. with active disease at the time of blood draw before surgery, 30 In a small number of patients with active disease (n ϭ 7), were diagnosed as having primary tumors, and 12 were diag- we were able to obtain blood samples before and 2–4 months nosed as having recurrent or second primary tumors (Table 1). after curative surgery. It was interesting to note that the absolute CD4ϩ and CD8ϩ T-cell counts were significantly depressed in count of CD4ϩ T cells decreased in four of seven patients after both these groups relative to NCs (Fig. 3B). Although a trend toward lower CD4ϩ T-cell counts was noted for patients with recurrence or second primary tumors, this difference did not reach statistical significance (data not shown). In addition, there were six patients with recurrent disease after previous surgery, and these subjects had the lowest T-cell counts (data not shown). Similarly, among the patients who, at the time of blood draw, had NED after curative therapy (n ϭ 98), those treated for recurrence of a second primary tumor (n ϭ 19) had, on average, a 25% lower number of CD4ϩ T cells than the patients treated for primary disease (n ϭ 79), as shown in Fig. 6 (P Ͻ 0.06). Neither disease stage, site, or nodal involvement defined for these patients at the time of curative surgery had any dis- cernible effects on absolute lymphocyte counts determined at the time of blood draws (data not shown). It has to be stressed, however, that blood draws for lymphocyte counts were obtained at different time points relative to surgery and, in many cases, 4 or more years after curative therapy.

DISCUSSION We sought to analyze both the percentages and absolute lymphocyte counts in a large cohort of SCCHN patients (n ϭ 146), 48 of whom had active disease, and 98 of whom had NED after having undergone curative therapy months or years before ϩ ϩ Fig. 5 Absolute counts of CD4 and CD8 T cells in individual the blood draws obtained for this study. The rationale for the patients with no evidence of disease (NED) after curative surgery. The patients studied at Ͻ2 years after surgery are represented by white study was solely based on our earlier reports of significant levels squares, whereas those studied Ͼ2 years after surgery are represented of spontaneous apoptosis observed among circulating T lym- by black squares. Note the low absolute counts of both CD4ϩ and phocytes in patients with SCCHN (2, 3). This finding suggested CD8ϩ T cells in the patients with NED, regardless of the length of that absolute lymphocyte counts might be depressed in these ϩ ϩ follow-up. No difference in the CD4 /CD8 ratio was observed be- patients. Through the use of single-platform flow cytometry, it tween these two groups of patients with NED (P ϭ 0.56). The stippled lines represent mean values for CD4ϩ and CD8ϩ T-cell number in was possible to independently analyze the percentages and normal donors. absolute numbers of the T-cell subsets in the patients’ peripheral

Fig. 6 Box plots showing absolute T-cell counts in squamous cell carcinoma of the head and neck patients with active disease at the time of blood draws (n ϭ 98; before surgery). Of these, 79 had surgery for a primary tumor. Nineteen patients had recurrent or second primary tumors. The Ps reflect comparisons between the two groups. The dashed horizontal reference lines represent the mean normal control values for CD4ϩ and CD8ϩ T-cell subsets.

circulation. Interestingly, absolute numbers of CD3ϩ, CD4ϩ, patients with head and neck cancer also bound annexin V (21). and CD8ϩ T cells were found to be significantly decreased in Thus, the homeostasis of CD8ϩ T cells in the peripheral circu- the patient group, emphasizing the existence of a relative lym- lation of patients with head and neck cancer appears to be phopenia in the patients versus age-adjusted NCs, despite the maintained by their rapid turnover, resulting in a relatively apparently normal percentages of circulating T lymphocytes. stable mean peripheral CD8ϩ cell count for the patient popu- Therefore, in this report, we emphasize the necessity of deter- lation as a whole. A situation associated with profound depres- mining absolute counts, not percentages, of lymphocyte subsets sion of the peripheral CD8ϩ T-cell count occurred in patients in patients with cancer. The usually reported cell percentages are with stage IV disease, who failed to normalize CD8ϩ T-cell misleading because they do not consider total WBC count, count after surgery. This observation suggests that in advanced which might be and frequently is altered in these patients, disease, patients might fail to effectively compensate for the loss particularly after anticancer therapies. of effector cells. The most interesting finding of this study, however, was Not surprisingly, the absolute counts of both CD8ϩ and the observation that the absolute count of CD8ϩ T cells, al- CD4ϩ T cells were found to be highly variable in the patient though significantly decreased in the patients versus NCs (and cohort we studied. For CD8ϩ T cells, the range varied from especially so in patients with active disease), appeared to re- Ͼ1500/mm3 to Ͻ100/mm3. For CD4ϩ T cells, the range was cover and normalize in most patients with NED, whereas the Ͼ2000/mm3 to Ͻ100/mm3. Patients with documented recur- CD4ϩ T-cell count did not. In view of the data reported by us rence of disease or second primary tumor at the time of blood previously, which had demonstrated preferential apoptosis of draw had the lowest T-cell counts of all studied. The observed CD8ϩ T cells in the circulation of SCCHN patients with clinical trend was for CD8ϩ T cell counts to increase and for CD4ϩ T characteristics similar to those of the patient cohort studied here cell counts to decrease after tumor removal by surgery. How- (2, 3), we had expected to find significantly decreased numbers ever, one of the most interesting findings of this study was that of circulating CD8ϩ T lymphocytes rather than CD4ϩ T lym- SCCHN patients who had NED for 2 years or Ͼ2 years after phocytes in these patients. A possible explanation for this un- curative surgery alone (no radiotherapy or chemotherapy) still expected finding could be that the homeostatic mechanisms showed an imbalance in the T-cell subsets. This imbalance was compensate for the selective apoptosis by a rapid expansion of highly significant for CD4ϩ T cells, whereas CD8ϩ T cells CD8ϩ T cells in the periphery of patients with SCCHN. It is tended to normalize. This observation suggests that the disease now known that lymphocyte numbers are not regulated through process has a profound and long-lasting impact on T lympho- control of their production from stem cells but rather by survival cyte homeostasis in patients with SCCHN. Others have sug- and homeostatic proliferation of mature lymphocytes (17, 18). gested previously that lymphocyte counts have prognostic value Thus, if the number of lymphocytes is low, proliferation ensues in SCCHN (22). and generates more cells. If the number of lymphocytes is high, Another factor that had a significant impact on T-cell excess lymphocytes die (17, 18). This is exactly what we have counts was the history of previous RT. Only CD4ϩ cells, but observed in the circulation of patients with SCCHN, where not CD8ϩ T cells, appeared to be affected by previous RT (Fig. expansion of CD8ϩ effector subsets (specifically, the 4). Whereas this has been reported previously (e.g., Ref. 23), the CD8ϩCD45ROϪCD27Ϫ or CD8ϩCD28Ϫ subsets) was ac- finding of decreased counts of CD4ϩ T cells long after the companied by their rapid demise by apoptosis (19, 20). More administration of RT was somewhat surprising. Apparently, recently, we have reported that the majority of CD4ϩ T cells are not only more sensitive to RT than CD8ϩ CD3ϩCD8ϩtetramerϩ T cells detectable in the circulation of lymphocytes, but the restoration of the peripheral CD4ϩ T cell

pool after RT is very slow. As indicated above, many patients REFERENCES with NED and no previous RT also had reduced CD4ϩ T-cell 1. Goldrath AW, Bevan MJ. Selecting and maintaining a diverse T-cell counts long after curative surgery. Thus, RT was not the only repertoire. Nature (Lond) 1999;402:255Ð62. factor modulating the number of circulating CD4ϩ cells in a 2. Dworacki G, Meidenbauer N, Kuss I, et al. Decreased ␨ chain cohort of patients with SCCHN previously treated for their expression and apoptosis in CD3ϩ peripheral T lymphocytes in patients disease. with melanoma. Clin Cancer Res 2001;7:947Ð57. It has been reported recently that lymphocyte counts are 3. Hoffmann T, Dworacki G, Tsukishiro T, et al. Spontaneous apoptosis of circulating T lymphocytes in patients with head and neck cancer and predictive of survival in cancer (11). Whereas the current study its clinical importance. Clin Cancer Res 2002;8:2553Ð62. was not designed to address survival as end point, we considered 4. Saito T, Kuss I, Dworacki G, et al. Spontaneous ex vivo apoptosis of the possibility that disturbed lymphocyte homeostasis might be peripheral blood mononuclear cells in patients with head and neck especially prominent in patients with poor prognosis, i.e., stage cancer. Clin Cancer Res 1999;5:1263Ð73. IV disease and nodal metastases. However, we could not estab- 5. Reichert TE, Strauss L, Wagner EM, Gooding W, Whiteside TL. lish a statistically significant relationship between disease stage Signaling abnormalities, apoptosis, and reduced proliferation of circu- or its severity and lymphocyte counts based on retrospective lation and tumor-infiltrating lymphocytes in patients with oral carcinoma. Clin Cancer Res 2002;8:3137Ð45. data analysis. It should be noted that no predictive value for 6. Hellerstein M, Hanley MB, Cesar D, et al. Directly measured kinetics absolute lymphocyte counts or CD4/CD8 ratios for larynx pres- of circulation T lymphocytes in normal and HIV-1-infected humans. Nat ervation, response to therapy, or survival of patients with SC- Med 1999;5:83Ð9. CHN was seen in another study (24). 7. Neese RA, Siler SQ, Cesar D, et al. Advances in the stable isotope- The observed imbalance in T lymphocyte counts in a mass spectrometric measurement of DNA synthesis and cell prolifera- heterogeneous cohort of patients with SCCHN is likely to be tion. Anal Biochem 2001;298:189Ð95. mediated by several distinct mechanisms. In patients with 8. Schroder W, Vering A, Stegmuller M, Strohmeier R. Lymphocyte subsets in patients with ovarian and breast cancer. Eur J Gynaecol Oncol NED, it could be due to a decreased thymic output of recent 1997;18:474Ð7. thymic emigrants, as reported by us previously (25), leading 9. Wolf GT, Amendola BE, Diaz R, et al. Definite vs. adjuvant radio- to alterations in the size of naõ¬ve, memory, and effector therapy. Comparative effects on lymphocyte subpopulations in patients pools. In patients with active primary or recurrent disease, it with head and neck squamous carcinoma. Arch Otolaryngol 1985;111: could be mediated by increased cell death (apoptosis) of 716Ð26. CD8ϩ T lymphocytes in the peripheral circulation (2, 3). We 10. Melichar B, Touskova M, Solichova D, Kralickova P, Kopecky G. ϩ are currently investigating each of these possibilities. CD4 T-lymphocytopenia and systemic immune activation in patients with primary and secondary liver tumours. Scand J Clin Lab Investig Clearly, the snapshot in time represented by each patient’s 2001;61:363Ð70. cell count, as reported here, does not reflect lymphocyte 11. Kay NE, Leong TL, Bone N, et al. Blood levels of immune cells kinetics within the circulating lymphocyte pool. Dynamic predict survival in myeloma patients: results of an Eastern Cooperative techniques, taking advantage of novel methods of cellular Oncology Group Phase 3 trial for newly diagnosed multiple myeloma labeling and serial measurements (6, 7), are necessary to patients. Blood 2001;98:23Ð8. confirm that the T-cell turnover is altered in patients with 12. Murta EF, de Andrade JM, Falcao RP, Bighetti S. Lymphocyte cancer relative to NCs. The use of CD4/CD8 ratios without subpopulations in patients with advanced breast cancer submitted to neoadjuvant chemotherapy. Tumori 2000;86:403Ð7. considering changes in absolute pool sizes can be misleading 13. Hernberg M, Muhonen T, Turunen JP, Hahka-Kemppinen M, Py- because changes in either the denominator or numerator rohonen S. The CD4ϩ/CD8ϩ ratio as a prognostic factor in patients might lead to the same ratio, as we have shown here. with metastatic melanoma receiving chemoimmunotherapy. J Clin On- Our results for patients with SCCHN are complementary col 1996;14:1690Ð6. to the findings of others who also report lymphocyte imbal- 14. Brando B, Barnett D, Janossy G, et al. Cytofluorometric methods ance in cancer (10Ð13). Combined with the presence of for assessing absolute numbers of cell subsets in blood. European functional abnormalities in T cells of patients with SCCHN, Working Group on Clinical Cell Analysis. Cytometry 2000;42:327Ð46. as reported by us previously (3, 5, 19, 20, 25), the overall 15. O’Gorman MR, Nicholson JK. Adoption of single-platform tech- impression is that of decreased immune competence in these nologies for enumeration of absolute T-lymphocyte subsets in peripheral blood. Clin Diag Lab Immunol 2000;7:333Ð5. patients. Lowered T-cell numbers in the circulation could 16. Reimann KA, O’Gorman MG, Spritzler J, et al. Multisite compar- predispose the patients to infections, disease recurrence, or a ison of CD4 and CD8 T-lymphocyte counting by single- versus multi- second malignancy. It is, therefore, advisable to pay attention ple-platform methodologies: evaluation of Beckman Coulter Flow- to T-cell counts during posttreatment visits, even when the Count fluorospheres and the tetraONE system. Clin Diag Lab Immunol patients have NED long after curative therapy. Our data 2000;7:344Ð51. emphasize that decreased T-cell counts reflect effects of the 17. Khaled AR, Durum SK. The role of cytokines in lymphocyte disease process on T-cell homeostasis and not just therapy- homeostasis. Biotechniques 2002;33:S40Ð5. mediated alterations and that the imbalance is long-lasting. In 18. Surh CD, Sprent J. Regulation of naõ¬ve and memory T-cell homeostasis. Microbes Infect 2002;4:51Ð6. view of the possibility that such persistent changes in homeo- 19. Kuss I, Donnenberg A, Gooding W, Whiteside TL. Effector stasis of T-cell subsets might adversely influence antitumor CD8ϩCD45ROϪCD27Ϫ T cells have signaling defects in patients with responses and promote recurrence, therapies designed to in- head and neck cancer. Br J Cancer 2003;88:223Ð30. crease lymphocyte counts could be considered even in pa- 20. Tsukishiro T, Donnenberg AD, Whiteside TL. Rapid turnover of the tients with NED. CD8ϩCD28Ϫ T-cell subset of effector cells in the circulation of pa-

tients with head and neck cancer. Cancer Immunol Immunother 2003; 23. Friedman EJ. Immune modulation by ionizing radiation and its impli- 52:599Ð607. cations for cancer immunotherapy. Curr Pharm Des 2002;8:1765Ð80. 21. Albers AE, Tsukishiro T, Ferris RL, DeLeo AB, WhitesideTL. TCR 24. Wolf GT, Bradford CR, Urba S, et al. Immune reactivity does not VB restrictions in peripheral circulation and tumor infiltrating lympho- predict immunotherapy response, organ preservation or survival in ad- cytes obtained from patients with head and neck cancer. Proc Am Assoc vanced laryngeal cancer. Laryngoscope 2002;112:1351Ð6. Cancer Res 2003;44:359. 25. Kuss I, Godfrey TE, Donnenberg AD, Whiteside TL. Low levels of 22. Wolf GT, Schmaltz S, Hudson J, et al. Alterations in T lympho- T-cell receptor excision circles (TREC) and paucity of naõ¬ve T cells in cyte subpopulations in patients with head and neck squamous the circulation of patients with cancer suggest a rapid lymphocyte carcinoma: correlations with prognosis. Arch Otolaryngol 1987;113: turnover within the memory compartment. Proc Am Assoc Cancer Res 1200Ð 6. 2002;43:278.

Downloaded from clincancerres.aacrjournals.org on September 30, 2021. © 2004 American Association for Cancer Research. Decreased Absolute Counts of T Lymphocyte Subsets and Their Relation to Disease in Squamous Cell Carcinoma of the Head and Neck

Iris Kuss, Bridget Hathaway, Robert L. Ferris, et al.

Clin Cancer Res 2004;10:3755-3762.

Updated version Access the most recent version of this article at: http://clincancerres.aacrjournals.org/content/10/11/3755

Cited articles This article cites 25 articles, 7 of which you can access for free at: http://clincancerres.aacrjournals.org/content/10/11/3755.full#ref-list-1

Citing articles This article has been cited by 20 HighWire-hosted articles. Access the articles at: http://clincancerres.aacrjournals.org/content/10/11/3755.full#related-urls

E-mail alerts Sign up to receive free email-alerts related to this article or journal.

Reprints and To order reprints of this article or to subscribe to the journal, contact the AACR Publications Subscriptions Department at [email protected].

Permissions To request permission to re-use all or part of this article, use this link http://clincancerres.aacrjournals.org/content/10/11/3755. Click on "Request Permissions" which will take you to the Copyright Clearance Center's (CCC) Rightslink site.