Multiple-dose Granulocyte-Macrophage–Colony-stimulating Factor Plus 23-valent Polysaccharide Pneumococcal Vaccine in Patients With Chronic Lymphocytic Leukemia

Abstract

BACKGROUND

For the current study, the authors sought to determine whether administration of multiple-dose granulocyte-macrophage–colony-stimulating factor (GM-CSF) could improve response to standard 23-valent polysaccharide pneumococcal vaccine (PPV) in patients with chronic lymphocytic leukemia (CLL).

METHODS

Patients were allocated randomly to receive PPV either alone or with 3 doses of GM-CSF (250 μg) given before or after vaccination. Serum was obtained before, 4 weeks after, and 12 weeks after vaccination for antibody determination. Thirty-two patients with CLL were given PPV. They were randomized to receive 3 doses of GM-CSF either before or after vaccination or to receive no GM-CSF.

RESULTS

A 4-fold rise in immunoglobulin G (IgG) to capsular polysaccharides from Streptococcus pneumoniae types 4, 6B, 9V, 14, 19F, and 23F occurred in <10% of patients in each of the 3 groups. There were no differences in geometric mean IgG levels in any of the 3 groups 4 weeks or 12 weeks after vaccination.

CONCLUSIONS

In patients with CLL, the response to pure polysaccharide pneumococcal vaccine was low despite immune enhancement with multiple doses of GM-CSF. In all patients, reactogenicity was minor.

Pneumococcal infections are serious complications in cancer patients with B-cell defects.¹ Chronic lymphocytic leukemia (CLL) is a malignancy of mature B-cells that most likely involves regulatory B-lymphocytes associated with the suppression of immunoglobulin G (IgG) production. The resulting hypogammaglobulinemia often is severe and lasting.² Patients with CLL who have low levels of antipneumococcal antibody are particularly at risk for severe and recurrent pneumococcal infections. Several preventive strategies have been used, including 1) vaccination against Streptococcus pneumoniae, 2) intravenous Ig therapy,³ and/or 3) prolonged prophylaxis with systemic antimicrobial agents. Antibody responses to standard pneumococcal vaccine are inadequate in most patients with B-cell cancers.⁴ Furthermore, a population-based efficacy analysis of 23-valent capsular polysaccharide pneumococcal vaccine (PPV) in an at-risk population was marginally effective in preventing pneumococcal bacteremia.⁵

Granulocyte-macrophage–colony-stimulating factor (GM-CSF) enhances multifaceted immune responses, including dendritic cell (DC)-mediated antigen presentation, and has been an attractive vaccine adjuvant.⁶ In patients with CLL, it is believed that abnormal cytokine profiles interfere with otherwise normal DC functions.^7,8 Adjuvant treatment with GM-CSF may enhance vaccine response in patients with CLL by promoting DC functions and/or normalizing B-cell responses.⁹ The objective of the current study was to determine whether administration of multiple-dose GM-CSF could improve response to pneumococcal vaccine in patients with CLL.

MATERIALS AND METHODS

Study Design

This study was approved by the Institutional Review Board of M. D. Anderson Cancer Center (MDACC) (Houston, Tex). Written informed consent was obtained from all patients. Patients were recruited from the adult CLL clinic at the MDACC between January 12, 2005 and July 12, 2006. They were allocated randomly to receive standard 23-valent PPV alone or with 3 doses of GM-CSF given before or after vaccination (Figure 1). The vaccine and GM-CSF were dispensed by the MDACC research pharmacy. Serum was obtained before, 4 weeks after, and 12 weeks after vaccination for antibody determination. Patients’ identifiable information was secured by creating a code, and all identifiers were secured at MDACC.

FIGURE 1.

Flow of patients with chronic lymphocytic leukemia through the study. PPV indicates polysaccharide pneumococcal vaccine; GM-CSF, granulocyte-macrophage–colony-stimulating factor.

Patients were eligible if they had biopsy-proven CLL according to standard diagnostic criteria.¹⁰ Patients were excluded if they had received fludarabine or alemtuzumab (monoclonal antibody that binds to CD52 and triggers antibody-dependent cell lyses) in the previous 12 months. In addition, patients who had received rituximab or parenteral Ig within 6 months were not included. Patients were ineligible if they were allergic to GM-CSF or PPV, if they had undergone splenectomy, if they had received systemic adrenal corticosteroids, or if they had been on other investigational vaccines or anti-neoplastic medications in the 2 weeks before enrollment. Because of eligibility criteria and the inconvenience of repeated visits, only 37 patients agreed to participate during the enrollment period.

Vaccine

PPV (Pneumovax; Merck & Company) was a commercially available pneumococcal vaccine from a single lot that contained purified capsular polysaccharides of 23 pneumococcal serotypes. Each dose contained 25 μg of each capsular polysaccharide and was administered as a single, 0.5-mL, intramuscular injection into the nondominant deltoid muscle.

GM-CSF

Sargramostim (Leukine; Berlex, Seattle, Wash) is a recombinant human GM-CSF that is produced by recombinant DNA technology in a Saccharomyces cerevisiae expression system. The recombinant product is a glycoprotein of 127 amino acids characterized by 3 primary molecular species, which have molecular masses of 19,500 days, 16,800 days, and 15,500 days. The amino-acid sequence of recombinant product differs from the natural human GM-CSF by a substitution of leucine at position 23, and the carbohydrate moiety may be different from the native protein. The biologic potency, as tested against the World Health Organization First International Reference Standard, is approximately 5.6 × 10⁶ IU/mg, and the dose used in this study was 250 μg.

Vaccine Response

Response to PPV was determined by measuring serum IgG to capsular polysaccharides from 6 of the most common infecting serotypes of S. pneumoniae using methods previously described from this laboratory⁵; the results are reported in μg/mL. An individual was regarded as having responded if there was a 2-fold rise in anticapsular IgG when the prevaccination titer was compared with levels either 4 weeks or 12 weeks postvaccination (whichever was higher) and with a final level >0.5 μg/mL. Geometric mean titers for each vaccine-dose group were calculated, and statistical comparisons were made using the Student t test. The rationale for the sample size was as follows: A sample size of 16 patients would have achieved 94% power to detect a difference of 50 between the null hypothesis mean of 250 and the alternative hypothesis mean of 200 with an estimated standard deviation of 60 and a 1-sided significance level of 5%. Therefore, 16 patients should have been sufficient to detect an inefficacious immune response.

Adverse Events

A diary card was provided to patients for self-documentation of adverse reactions that included soreness, tenderness, erythema, or swelling at the vaccination site; malaise, myalgia, feeling feverish, and actual fever were also recorded. Safety follow-up visits were at 7 days to 10 days, 28 ± 4 days, and 56 ± 4 days after receiving vaccination.

RESULTS

Study Population

Thirty-two individuals completed the protocol and were included in the final results. The patients and disease characteristics are shown in Table 1. Among 4 patients with hypogammaglobulinemia, IgG levels were 325 mg/dL, 265 mg/dL, 160 mg/dL, and 110 mg/dL. Rai-Sawitsky stages were comparable among the study groups: Of the patients who received GM-CSF before vaccination, 46% (n = 6) were in complete clinical remission (CR), and 54% (n = 7) had stable disease. Among the patients who received GM-CSF after vaccination, 58% (n = 7) were in CR, and 42% (n = 5) had stable disease. Among 7 patients who received vaccine alone, most patients (71% n = 5) had cancer in CR.

TABLE 1.

Patient and Underlying Disease Characteristics

Characteristic	No. of patients (%)
No. of patients	32
Sex
Men	17 (53)
Age: Median ± SD, y	68 ± 10
Race
White	28 (88)
Hispanic	1 (3)
Black	3 (9)
White blood cell count: Median ± SD, cells/UL*	6000 ± 23,861
Absolute neutrophil count: Median ± SD, cells/UL†	2230 ± 1407
Absolute lymphocyte count: Median ± SD, cells/UL‡	1455 ± 21,973
Ig levels prior to vaccination: Median ± SD, mg/dL
IgG	718 ± 665
IgA	86 ± 71
IgM	33 ± 55
Hypogammaglobulinemia§	4 (13)
Patients with CLL	32 (100)
Complete remission	18 (56)
Stable disease	13 (41)
Partial remission	1 (3)
CLL Rai-Sawitsky stage||	14 (44)
0	10 (31)
I	2 (6)
II	1 (3)
IV	1 (3)
Chemotherapy prior to vaccination	28 (87)
Rituximab-based antineoplastic regimen	23 (72)
Last chemotherapy dose and vaccination: Median ± SD, d	420 ± 666
Last rituximab dose and vaccination: Median ± SD, d	377 ± 416

SD indicates standard deviation; CLL indicates chronic lymphocytic leukemia, Ig, immunoglobulin.

^*White blood cell count within 1 week prior to vaccination.

^†Absolute neutrophil count within 1 week prior to vaccination.

^‡Absolute lymphocyte count within 1 week prior to vaccination.

^§Hypogammaglobulinemia was defined as IgG levels <350 mg/dL; all Ig levels were obtained within 1 week prior to vaccination.

^||Patients in the group had either stable disease or CLL in partial remission.

Prior antineoplastic therapy had consisted of fludarabine, cyclophosphamide, plus rituximab (murine/human monoclonal antibody with high affinity for CD20-expressing cells) in 14 patients (44%). Nine patients (28%) had received rituximab alone. One patient each had received fludarabine plus cyclophosphamide; fludarabine plus cyclophosphamide and amifostine; clofarabine; fludarabine; and a research plasmid vaccine. Four patients (13%) had not received antineoplastic therapy.

Vaccine Response

The number of individuals with an increase in IgG to capsular polysaccharides and the geometric mean serum antibody titer for each group in relation to vaccine and GM-CSF status are shown in Table 2. Of all possible individual responses (6 capsular polysaccharides × the number of patients in each group), only 8 of 78 patients (10.3%), 5 of 60 patients (8.3%) and 4 of 42 patients (9.5%) were positive in the GM-CSF then PPV23 group, the PPV23 then GM-CSF group, and the PPV23 alone group, respectively. One patient who received GM-CSF before PPV23 had received weekly rituximab (1085 mg) during 4 weeks after vaccination. In this patient, no difference in antipneumococcal IgG titers was noted before or after vaccination. After excluding this patient from the final analysis shown in Table 2, the values of geometric mean titer did not change significantly (data not shown).

TABLE 2.

Serum Antibody Responses to 23-Valent Polysaccharide Pneumococcal Vaccine in Patients With Chronic Lymphocytic Leukemia

Pneumococcal Serotype	GM-CSF, then PPV23		PPV23, then GM-CSF		PPV23 alone
	No. with increase/total No. (%)*	Mean PoV/PreV [Fold increase]	No. with increase/total no. (%)*	Mean PoV/PreV [Fold increase]	No. with increase/total no. (%)*	Mean PoV/PreV [Fold increase]
4	0/13 (0)	0.39/0.27 [1.4]	0 (0)	0.23/0.16 [1.4]	1/7 (14)	0.45/0.31 [1.4]
6B	1/13 (8)	0.77/0.53 [1.5]	1/12 (8)	0.48/0.24 [2]	0 (0)	1.92/1.69 [1.1]
9V	1/13 (8)	0.55/0.30 [1.8]	1/12 (8)	0.56/0.45 [1.2]	1/7 (14)	1.48/0.84 [1.8]
14	2/13 (15)	4.06/2.01 [2]	2/12 (17)	1.33/0.91 [1.5]	0 (0)	4.38/3.13 [1.4]
19F	2/13 (15)	1.42/1.12 [1.3]	0 (0)	0.93/0.98 [0.9]	1/7 (14)	1.56/0.73 [2.1]
23F	2/13 (15)	0.42/0.47 [0.9]	1/12 (8)	0.77/0.62 [1.2]	1/7 (14)	0.61/0.39 [1.6]

GM-CSF indicates granulocyte-macrophage–colony stimulating factor; PPV23, 23-valent polysaccharide pneumococcal vaccine; PoV, the highest geometric mean titer (GMT) in μg/mL 4 weeks and 12 weeks after vaccination; PreV, prevaccination and pre-GM-CSF GMT.

^*The number of patients who had an increase ≥2-fold and a final immunoglobulin G level >0.5 μg/mL/the total number of patients.

Comparing prevaccine and postvaccine IgG levels for each individual polysaccharide and within each group, in no patient was there a statistically significant difference (P >.05 for all comparisons; Student t test). A 2-fold increase in geometric mean IgG levels was observed for only 1 capsular polysaccharide in each group, and the overall rises in titers were remarkably similar and equally low (differences were not significant for all comparisons; P >.05).

Vaccine-related Adverse Events

All patients were followed for 6 months after vaccination, and no serious vaccine-related adverse reactions were noted. Among patients who received GM-CSF before vaccination, 1 patient each had pain and redness at the cytokine injection site, palmer erythema, and a low-grade fever (38 °C) for 1 day. Whereas, in patients who received GM-CSF after PPV, 5 patients had pain and redness at the injection site, and 1 patient developed self-limiting cold-like symptoms. In patients who received PPV alone, 1 patient developed pain and redness at the injection site.

DISCUSSION

In this randomized study of patients with CLL, multiple-dose GM-CSF given before or after PPV was not associated with enhanced serum antibody responses to pneumococcal capsular polysaccharides compared with responses in patients who were given PPV alone. The use of 3 GM-CSF doses before or after PPV was safe and produced no serious adverse reactions in any patients.

The use of GM-CSF as a single dose with standard, trivalent influenza vaccine was ineffective in improving vaccine response in patients with solid organ cancers or hematologic malignancies, including patients who had received adjuvant chemotherapy within 6 weeks of vaccination.¹¹ However, a single dose of GM-CSF improved the response to influenza vaccine in stem cell transplantation recipients during the early post-transplantation period.¹² In our CLL patients, even multiple GM-CSF doses before or after pneumococcal vaccination did not improve antibody responses at 4 weeks or 12 weeks after vaccination (Table 2).

A low immunogenicity of PPV, which predominantly depends on the IgG2-mediated immune response, may be the reason for the poor performance of PPV in patients with clonal B-cell leukemia who have selective IgG subclass defects. Other factors also may have played a role because advanced disease, low gammaglobulin levels, and high soluble CD23 blood levels were factors associated with unfavorable vaccine response in 25 patients with CLL.¹³ The results of the current study negate our hypothesis that impaired antibody responses to PPV among patients with CLL may be improved by administering multiple doses of GM-CSF.

Various strategies to improve vaccine responses have yielded inconsistent results in patients with B-cell malignancies. Recently, we demonstrated a benefit of a 6-fold higher than standard influenza vaccine dose given to patients with B-cell non-Hodgkin lymphoma.¹⁴ In another study, immune priming with a peptide-conjugate pneumococcal vaccine resulted in enhanced antipneumococcal antibody response to subsequent vaccination with PPV.¹⁵ Further studies are needed to determine the value of an approach that uses 1 or more method(s) to improve vaccine-response in immunosuppressed patients with hematologic malignancies. Such studies are ongoing in patients with CLL.