Rituximab-induced T cell depletion in patients with rheumatoid arthritis: association with clinical response.

OBJECTIVE: Rituximab, a monoclonal antibody specifically targeting CD20, induces B cell depletion and is effective in the treatment of rheumatoid arthritis (RA). This study was undertaken to evaluate whether routine monitoring of lymphocyte subpopulations, especially T cells, may be useful in patients receiving rituximab for RA. METHODS: We examined data on all RA patients receiving rituximab between July 2007 and November 2012 in our center. Peripheral blood CD3+, CD4+, CD8+, CD3-CD56+, and CD19+ lymphocyte counts before and during the first course of rituximab were measured by flow cytometry. The Mann-Whitney nonparametric test was used to compare lymphocyte subpopulation counts before and during treatment. RESULTS: Data on 52 patients were examined. Rituximab induced unexpected and substantial depletion of T cells, mainly CD4+ cells, in most patients. The CD4+ cell count decreased by a mean ± SD of 37 ± 33% as compared to baseline at week 12, reaching <200 cells/µl in 3 patients. Importantly, lack of CD4+ cell depletion was associated with no clinical response. Therefore, the mechanism of action of rituximab may depend at least in part on T cells. CONCLUSION: Rituximab induces substantial T cell depletion, mainly of CD4+ cells, which is associated with the clinical response in RA. Routine monitoring of T cells may be useful in the clinical setting of RA.

Efficiency of immunoglobulin G replacement therapy in common variable immunodeficiency: correlations with clinical phenotype and polymorphism of the neonatal Fc receptor.

Treatment of common variable immunodeficiency disorders (CVID) is based on replacement therapy using intravenous (i.v.) or subcutaneous (s.c.) immunoglobulin (Ig)G. Interindividual variation of IgG dose is common. A total of 380 CVID patients on stable IgG replacement from two prospective cohorts were analysed. An 'efficiency' index was defined as the ratio of serum IgG trough level minus IgG residual to the average weekly dose of IgG infusion. A reduced efficiency of IgG was associated independently with the i.v. route (P < 0·001) and with the presence of at least one CVID disease-related phenotype (lymphoproliferation, autoimmune cytopenia or enteropathy) (P < 0·001). High IgG efficiency was noted in patients homozygotes for the variable number tandem repeat (VNTR) 3/3 polymorphism of the neonatal Fc receptor gene [IgG Fc fragment receptor transporter alpha chain (FCGRT)] promoter, and this was particularly significant in patients treated with IVIG (P < 0.01). In a multivariate analysis, FCGRT VNTR 3/3 genotype (P = 0·008) and high serum albumin (P < 0·001) were associated independently with increased efficiency of i.v. Ig.

Association of Fcγ receptor IIIA variant with a subset of anti-topoisomerase I-positive patients in systemic sclerosis: a descriptive pilot study.

Hypothesizing a pathophysiological role of anti-topoisomerase I antibodies (anti-topo I) through autoantibody-dependent cell-mediated cytotoxicity (ADCC) and cytotoxic effectors expressing receptors for the Fc portion of IgG in systemic sclerosis (SSc), 267 SSc patients (56 with anti-topo I and 102 with anti-centromere antibodies (ACA)) were genotyped for the functional FCGR3A-V158F polymorphism. A descriptive analysis of patients according to their clinical and immunological status and FCGR3A-158 V/F genotypes was performed using multiple correspondence analysis. This descriptive analysis revealed an association between the FCGR3A-158 VV genotype and the presence of anti-topo I. By contrast, no relationship was found between FCGR3A polymorphism and the presence of ACA. SSc patients with anti-topo I appear to be more frequently homozygous for the high-affinity FcγRIIIA-coding allele, suggesting that some autoantibodies may be pathogenic through ADCC.

Neutrophil role in in vivo anti-lymphoma activity of rituximab: FCGR3B-NA1/NA2 polymorphism does not influence response and survival after rituximab treatment.

BACKGROUND: Neutrophils could play an important role in in vivo rituximab anti-lymphoma activity. FcgammaRIIIb is expressed only by neutrophils and FcgammaRIIIb-neutrophil antigen (NA)1/NA2 polymorphism influenced phagocytosis of immunoglobulin G1-opsonized particles. We formulated the hypothesis that if neutrophils are critical cells for in vivo rituximab activity, FcgammaRIIIb-NA1/NA2 polymorphism could influence the response to rituximab. PATIENTS AND METHODS: FCGR3B-NA1/NA2 genotypes were determined in 46 patients having received rituximab for a previously untreated, follicular, non-Hodgkin's lymphoma. The clinical response and the disappearance of the BCL2-JH gene rearrangement in both peripheral blood and bone marrow were evaluated at 2 months (M2) and each year during 7 years. RESULTS: They were 13% homozygous for FCGR3B-NA1, 61% homozygous for FCGR3B-NA1/NA2 and 26% heterozygous. The objective response rates at M2 were 67% in homozygous FCGR3B-NA1 patients compared with 75% in homozygous FCGR3B-NA2 and 75% in heterozygous patients (not significant). We found no difference for progression-free and overall survival by FCGR3B-NA1/NA2 genotypes. CONCLUSION: These results indicate no association between FCGR3B-NA1/NA2 polymorphism and response to rituximab indicating no significant role of phagocytosis mediated by neutrophils in in vivo mechanism of rituximab activity.

[Molecular pharmacogenetics in hospital laboratories in France: current data and future prospects]. / La pharmacogénétique moléculaire hospitalière en France: données actuelles et perspectives.

Molecular pharmacogenetic units have recently been established in several hospital laboratories in France. The clinical impact of these units is still limited and numerous problems of organizational, ethical, legal, technical, social and economical nature remain to be resolved. However, an increasing number of these units, a rise in their activities and an enlargement of their scope of application are foreseeable in the future. Ultimately, these units would significantly contribute to limit the public health problem caused by interindividual variabilities in drug effects. In view of these prospects, it seems essential that such hospital activity should be quickly recognised by the authorities and the various health sectors in France. It is also essential that the problems that arise from such pharmacogenetic activities should be considered by the authorities and would profit from the organization of a national network and from financial guarantees.

[Find your way in the jungle of mAbs]. / Comment se repérer dans la jungle des mAbs ?

The rapidly increasing number of approved monoclonal antibodies (mAbs) and the huge number of mAbs in clinical development are a matter of concern for who wants to easily identify targets, indications, mechanisms of action and possible adverse effects. The current nomenclature being of limited interest, simple rationales will be presented for helping practitioners in rapidly classify mAbs depending on their structure-pharmacology relationship and in evaluating their potential effects, particularly in transfusion medicine.

ELISA for the detection of anticardiolipin antibodies. High specificity based on the use of adult bovine serum as buffer and systematic subtraction of non-specific binding.

We have used an ELISA procedure to compare the reactivity of samples with or without IgG anticardiolipin antibodies (ACA) on cardiolipin-coated wells (target wells) and cardiolipin-free wells (control wells), using 10% fetal calf serum (10% FCS), 10% adult bovine serum (10% ABS) or 1% bovine serum albumin (1% BSA) as buffer. With 1% BSA, ACA reactivity was very low which suggests that this buffer would be inappropriate for use in ELISA procedures for ACA. 10% FCS induced non-specific binding of normal IgG only on target wells, particularly in the case of IgG hypergammaglobulinemia. With 10% ABS, this non-specific binding occurred on the solid phase with and without cardiolipin and could be accurately evaluated by subtracting the absorbance of control wells. Results of assays on 35 systemic lupus erythematosus sera using 10% ABS as buffer showed that highly specific results-could be obtained only if non-specific binding was systematically subtracted.

Evidence of noninvolvement of swine MHC class II in the in vitro proliferative response of human lymphocytes to porcine endothelial cells.

Successful pig-to-human xenotransplantation may expose swine endothelium to the human immune system. Since endothelial MHC class II expression is crucial in the genesis of an allogeneic lymphocyte response, the involvement of porcine MHC (SLA) class II molecules in the induction of human lymphocyte proliferation was studied. When cocultured with a confluent monolayer of irradiated porcine aortic endothelial cells (PAEC), human peripheral blood mononuclear cells (PBMC) incorporated tritiated thymidine. Monocyte depletion strongly reduced the magnitude of the lymphocyte proliferative response. Resting cultured PAEC were SLA class II-negative and an induction of these molecules during the xenogeneic mixed lymphocyte endothelial cell culture (XMLEC) was not observed. Moreover, the addition of an antibody directed against the SLA-DR molecule was without effect. Lymphocyte proliferation was also studied in response to SLA class II-positive stimulating cells--either human TNF-alpha-stimulated PAEC or porcine splenocytes. Induction of SLA class II molecules on PAEC had no effect on the human PBMC proliferative response. Moreover, human PBMC did not proliferate in response to porcine splenocytes. These results suggest (1) that SLA class II molecules are not involved in the induction of the human lymphocyte proliferative response and (2) that the endothelial nature of the stimulating cells plays a key role in this proliferation.

Removal of terminal alpha-galactosyl residues from xenogeneic porcine endothelial cells. Decrease in complement-mediated cytotoxicity but persistence of IgG1-mediated antibody-dependent cell-mediated cytotoxicity.

To determine the role of the terminal alpha-galactosyl residue in the endothelial damage mediated by human xenoreactive natural antibodies (IgM and IgG), we treated porcine endothelial cells in culture with green coffee bean alpha-galactosidase. A practically complete removal of terminal alpha-Gal residues (as evaluated by flow cytometry with Bandeiraea simplicifolia isolectin B4) and concomitant exposure of N-acetyllactosamine were obtained without altering cell viability. A dramatic decrease in IgM and IgG binding (from a pool of human sera) was observed, confirming the key role of the alpha-galactosyl residues. The enzyme treatment did not induce any nonspecific immunoglobulin binding sites, but led to the exposure of new epitopes for a minor fraction of IgM. The main residual IgM and IgG binding could be due to xenoantigens other than the alpha-galactosyl residues. When alpha-galactosidase-treated endothelial cells were used as targets in cytotoxicity experiments, they were less susceptible than untreated cells to complement-mediated cytotoxicity induced by fresh human serum. In contrast, they did not acquire resistance to human IgG-dependent cellular cytotoxicity, despite the decrease in IgG binding. Because it is known that antibody-dependent cytotoxicity mediated by CD16+ NK cells is dependent on IgG1 and IgG3, and not on IgG2 or IgG4, which was confirmed by blocking experiments, we studied the binding of all four subclasses to intact and alpha-galactosidase-treated endothelial cells. Two major subclasses, IgG1 and IgG2, bound to untreated endothelial cells, whereas IgG3 binding was low and IgG4 binding was negligible. A decrease in IgG1, IgG2, and IgG3 binding was observed upon alpha-galactosidase treatment, indicating that antibodies belonging to these three subclasses recognize alpha-galactosyl residues. The decrease in IgG2 binding was more pronounced than the decrease in IgG1 binding. Collectively, these data indicate that IgG1 xenoreactive natural antibodies, including those which are not directed at the alpha-galactosyl residues, could play a major role in the early delayed vascular rejection of pig xenografts.

Tissue factor activity of syncytiotrophoblast plasma membranes and tumoral trophoblast cells in culture.

During pregnancy, important modifications of hemostasis occur resulting in mothers in hypercoagulability and the role of placental cells such as trophoblast cells has been hypothesized. In this study, we first showed that syncytiotrophoblast plasma membranes, isolated from normal human placenta, expressed a strong tissue factor (TF) activity. We then studied TF activity of two continuous trophoblast cell lines (JEG-3 and BeWo) in comparison to human umbilical vein endothelial cells (HUVEC) and transformed human endothelial cells (ECV-304). TF assays were performed on intact detached confluent cells. Unstimulated JEG-3 and BeWo cells exhibited a very high TF activity which slightly increased after 2 to 4 h TNF-alpha stimulation. In contrast, HUVEC and ECV-304 had a lower basal TF activity which was mainly inducible by TNF-alpha, with a maximum effect after 4 to 6 h stimulation. For both cell types, TF activity was decreased to basal value after 16-hour TNF-alpha stimulation. These results support that trophoblast cells are able to express TF but the involvement of this property in the hemostatic physiological changes observed during pregnancy, remains to be demonstrated.

Association between polymorphism in IgG Fc receptor IIIa coding gene and biological response to infliximab in Crohn's disease.

AIM: To test the hypothesis of an association between polymorphism in FCGR3A (the gene coding for FcgammaRIIIa, which is expressed on macrophages and natural killer cells, is involved in antibody-dependent cell-mediated cytotoxicity and has recently been associated with a positive response to rituximab, a recombinant immunoglobulin G1 antibody used in non-Hodgkin's lymphomas) and response to infliximab in Crohn's disease. METHODS: FCGR3A-158 polymorphism was determined using an allele-specific polymerase chain reaction assay in 200 Crohn's disease patients who had received infliximab for either refractory luminal (n = 142) or fistulizing (n = 58) Crohn's disease. Clinical and biological responses (according to C-reactive protein levels) were assessed in 200 and 145 patients, respectively. RESULTS: There were 82.9% clinical responders in V/V patients vs. 72.7% in V/F and F/F patients (N.S.). Globally, the decrease in C-reactive protein was significantly higher in V/V patients than in F carriers (P = 0.0078). A biological response was observed in 100% of V/V patients, compared with 69.8% of F carriers (P = 0.0002; relative risk, 1.43; 95% confidence interval, 1.27-1.61). In the sub-group of patients with elevated C-reactive protein before treatment, the multivariate analysis selected the use of immunosuppressive drugs and FCGR3A genotype as independent factors influencing the clinical response to infliximab (P = 0.003). CONCLUSION: Crohn's disease patients with FCGR3A-158 V/V genotype have a better biological and, possibly, clinical response to infliximab.

Human NK cell-mediated direct and IgG-dependent cytotoxicity against xenogeneic porcine endothelial cells.

Once hyperacute rejection has been prevented, the pig-to-human xenograft might be exposed to vascular cell-mediated rejection directed against vascular structures. In order to evaluate the relative importance of direct and antibody-dependent anti-endothelial cell-mediated cytotoxicity in different individuals, freshly isolated human blood leukocytes were incubated with confluent porcine aortic endothelial cells (PAEC) in a 4 h Cr-release cytotoxicity assay. Peripheral blood mononuclear cells (PBMC) and lymphocytes (PBL) of all subjects tested (but not monocytes or neutrophils) directly killed PAEC, with wide interindividual variations (from 2.8% to 32%). The addition of heat-inactivated autologous serum to PBMC and PBL (but not to myeloid cells) always enhanced cytotoxicity. This antibody-dependent cell-mediated cytotoxicity (ADCC) was also observed in the presence of adult pooled serum and cord blood pooled serum and was eliminated after adsorption of adult pooled serum to insoluble protein A, demonstrating that IgG is the only class of immunoglobulin involved in this phenomenon. Moreover, blocking Fc gamma RIII with an anti-CD16 mAb eliminated ADCC without affecting direct cytotoxicity. When the ADCC exerted by the PBL of all subjects was assessed with the same preparation of purified IgG, wide interindividual variations were again observed. Surprisingly, there was no correlation between direct cytotoxicity and ADCC although, as depletion experiments demonstrated, both were due to CD16+ natural killer (NK) cells. These results argue that CD16+ NK cells could play an important role in early vascular rejection of porcine discordant xenografts, by both a direct and an IgG xenoreactive natural antibody-dependent cell-mediated cytotoxicity.

Development of an ex vivo model of pig kidney perfused with human lymphocytes. Analysis of xenogeneic cellular reactions.

BACKGROUND: Because of the explosive nature and the extremely rapid process of hyperacute rejection (HAR), significant infiltration of the xenograft by immunocompetent cells is not observed, and the role and the mechanism of action of cell-mediated rejection in discordant xenografts are therefore still under discussion. METHOD: We developed an experimental approach using pig kidneys perfused with human peripheral blood lymphocytes (PBL) in which the immunologic barrier of hyperacute rejection was excluded and which mimics the in vivo situation. RESULTS: PBL retention in the kidney was evaluated at 20-minute intervals for 3 hours. Retention increased from 30% to 80% with the time of perfusion and was specific because significantly fewer syngeneic lymphocytes were retained. Phenotype analysis of recovered PBL showed a significant decrease in natural killer (NK) cells. Immunohistochemical studies revealed the presence of NK cells and T lymphocytes in the glomerular and interstitial tubular structures of the kidney. Functional studies showed a progressive cessation of diuresis and augmentation of renal vascular resistance when the kidney was perfused with PBL. Electron microscopy examinations of kidney sections perfused with PBL showed swollen endothelial zones, suggesting alterations to and damage of the endothelium. CONCLUSIONS: This system provides a valuable model for the study of early discordant xenogeneic cellular rejection and demonstrates the predominance of xenograft infiltration by NK cells.

Removal of primate xenoreactive natural antibodies by extracorporeal perfusion of pig kidneys and livers.

Organ perfusion is one of the possible strategies to attenuate rejection of discordant xenografts by reducing the levels of the recipient's xenoreactive natural antibodies (XNA). Its efficacy in terms of XNA removal was studied in models of primate blood or plasma perfusion through porcine kidneys or livers, with special attention to haematological consequences and potential side-effects. We first perfused the blood of rhesus monkeys through pig kidneys and livers, and demonstrated that the perfusion of a pig liver resulted in higher XNA adsorption (72 +/- 13%) than the perfusion of a pig kidney (51 +/- 25%). However, when we normalized for the weight of the perfused organs and for levels of natural antibodies in individual monkeys, livers adsorbed less antibody (1.4 +/- 0.9 U antibody/g) than kidneys (7.2 +/- 7 U antibody/g). Histological signs of rejection were observed in perfused kidneys, but not in perfused livers. A major drawback of the perfusion of blood through livers was a considerable decrease in the primates' haemoglobin and platelet levels. To avoid this, we developed a plasma liver perfusion device. This method allowed a significant improvement in the haemodynamic state of primates and was particularly effective in preventing anaemia. Moreover, plasma liver perfusion was as effective as blood liver perfusion to remove natural antibodies and, resulted in a marked decrease in their functional activity as assessed by complement-dependent cytotoxicity (CDC) and antibody-dependent cell-mediated cytotoxicity (ADCC). The level of other plasma proteins was not significantly affected, apart from a dilution effect. After xenoperfusion a strong antibody response was evidenced by ELISA, CDC and ADCC between days 7 and 14 and then decreased progressively. We conclude that the separation of blood to allow the perfusion of plasma through a pig organ is safer than the perfusion of unseparated blood and is associated with efficient natural antibody removal. However, organ perfusion is limited by a rebound in antibody levels after a few days, and thus will have to be associated with anti-B cell immunosuppressive therapy for long-term or repeated applications.

A simplified and low-cost one-stage chromogenic assay for tissue factor dependent procoagulant activity of endothelial cells.

We developed a sensitive tissue factor (TF) chromogenic assay on a limited number of endothelial cells (EC), performed in microtiter plates, and which uses a normal pooled human plasma instead of purified concentrates as a source of coagulation factors. Primary cultures of human umbilical vein EC (HUVEC), both unstimulated and stimulated by lipopolysaccharide (LPS) were incubated with 50 microliters of of diluted normal human plasma (NHP) and 50 microliters of Factor Xa-specific chromogenic substrate (CBS 31-39, Stago, France). Hirudin was added at 4 U/ml to the plasma/CBS 31-39 mixture to inhibit thrombin generation. Optical densities were read at 405 nm and corresponding amounts of generated factor Xa were expressed in mU Xa/well using a standard curve established with purified human Factor Xa. The following parameters were then defined: the number of EC to plate (10(4) EC/well of a 96-well plate), the plasma-test dilution (1:20), the concentration of CBS 31-39 (0.50 mM) and the incubation time of reagents with EC (2 hours). The procoagulant activity (PCA) measured was only dependent on TF since it was no longer detectable either when FVII-deficient plasma was tested instead of normal human plasma or when PCA assays were performed in the presence of a blocking anti-human TF monoclonal antibody. This method allowed detection of a TF-dependent PCA on as few as 1000 EC per well. In addition, TF expression equal to 50% of maximal values was measured with LPS concentrations as low as 1 ng/ml, supporting the high sensitivity of the assay.

Effect of rIFN-gamma and IL-2 treatments in mouse and nude rat infections with Toxoplasma gondii.

Mice and nude rats lethally infected with T. gondii and treated with recombinant rat interferon-gamma (rIFN-gamma) or recombinant human interleukin-2 (rIL-2) were protected against death, when compared with untreated infected controls. In mice rIFN-gamma and rIL-2 played an important role in "prophylactic treatment", but not in "curative therapy". The survival rate was 42% in mice treated with 3 doses of 20,000 U of rIFN-gamma at days -2, -1, 0 before challenge and up to 66% in mice treated with 3 doses of 10,000 U of rIFN-gamma at days -2, 0, +2 before and after infection. Whereas the survival rate was 33% in mice that received 3 doses of 500 U rIL-2 at days -2, -1, 0 before infection, or -2, 0, +2 before and after infection respectively, up to 50% of the mice treated with 3 doses of 1,000 U rIL-2 at days -2, -1, 0 survived. In nude rats rIFN-gamma had a slight effect in "prophylactic treatment", whereas rIL-2 was active only in "curative treatment". The survival rate was 25% both in nude rats treated with doses of 400,000 U of rIFN-gamma at days -3, 0 before challenge, or with doses of 5,000 U of rIL-2 at days +2, +6, +9 after infection. These results lead us to hypothesise that the mechanism by which the lymphokine treatment exerts a protective effect on Toxoplasma infected mice is different from that on nu/nu rats. We conclude that these cytokines may play a notable role in modulating the host's immune defence against T. gondii infection.

Docosahexaenoic and eicosapentaenoic acids inhibit in vitro human endothelial cell production of interleukin-6.

The interaction between lymphocytes, cytokines, and endothelial cells (EC) is a key step in the inflammatory process. Interleukin-6 (IL-6) a pleiotropic cytokine in its effects, seems to be an early indicator of acute systemic inflammation. In this study, we have examined the effects of polyunsaturated fatty acids (PUFAs) on the production of IL-6 by human unstimulated EC or EC stimulated with TNF-alpha (100 U/ml); IL-4 (100 U/ml); LPS (1 ug/ml); or allogeneic peripheral blood lymphocytes (PBL). Twenty-four hour culture supernatants of immunoreactive IL-6 were measured by Sandwich ELISA. We have shown that the production of IL-6 was potentiated when EC were stimulated with TNF-alpha; IL-4; LPS; or monocyte-depleted PBL in comparison to unstimulated EC. The addition of n-3 PUFAs in culture medium (100 ug/ml DHA or EPA) significantly reduces the production of IL-6 by unstimulated EC; or stimulated with TNF-alpha; IL-4 pg/ml); LPS or depleted PBL respectively for DHA and EPA, whereas the n-6 PUFAs (Arachidonic acid), even used at the highest concentration, was ineffective. This inhibitory effect is PUFA dose dependent but is more potent with EPA than DHA. Regardless of the mode of action, since IL-6 is known to be involved in hematopoiesis, in the regulation of the immune response and in the inflammatory reaction, these results suggest that n-3 PUFAs may play a role in suppressing inflammation. Further studies are needed to elucidate the mechanism involved and the choice between the two fatty acids for clinical and therapeutic purposes.

[Role of the endothelium in leukocyte trafficking]. / Rôle de l'endothélium dans le trafic leucocytaire.

Vascular endothelium controls leukocyte recirculation and recruitment, receiving messages from underlying tissues and transforming them into informations quickly read by the passaging leukocytes. These messages induce adhesive interactions between leukocytes and endothelium via adhesion molecules. Multiple receptor-ligand pairs function in 4 steps: initial tethering and rolling of leukocytes; leukocyte triggering by chemotactic factors; arrest and strong adhesion to endothelium; and ultimately transendothelial migration (diapedesis).

Pre-treatment lymphopenia as a prognostic biomarker in colorectal cancer patients receiving chemotherapy.

BACKGROUND: Lymphopenia is a predictor of the efficacy and hematological toxicity of chemotherapy in various advanced cancers. There is little data about this relationship in colorectal cancer. In this retrospective study, the influence of pretreatment lymphopenia on hematological toxicity and the efficacy of chemotherapy was investigated in colorectal cancer patients. PATIENTS AND METHODS: In total, 260 patients were included in the study. Correlations between pre-treatment lymphopenia (lymphocyte count < 1,000/µl) and the occurrence of hematological toxicity and efficacy of first-line palliative chemotherapy were investigated. RESULTS: Lymphopenia was found in 49/260 (19%) patients. Ten of these patients with lymphopenia (20.4%) experienced severe hematological toxicity compared with 17 of the remaining 211 (8%) patients (P = 0.01). Lymphopenia was identified as an independent factor for hematological toxicity. Among patients who received palliative chemotherapy, the objective response rate was significantly lower in lymphopenic patients than in the other patients (12.5% vs. 40.2%; P = 0.004). Lymphopenia was strongly associated with shorter progression-free survival (median 4 vs. 7 months; P = 0.033) and shorter overall survival (median 16 vs. 24 months, P = 0.024). Multivariate analysis revealed that lymphopenia had an independent effect on survival. CONCLUSIONS: Our findings show that lymphopenia is an independent predictive factor for both hematological toxicity and efficacy of chemotherapy in colorectal cancer. Pre-treatment lymphocyte count may represent a simple and new predictive biomarker of chemotherapy effects in colorectal cancer patients.