Article Text

Download PDFPDF

Tumour-infiltrating macrophages and clinical outcome in breast cancer
  1. S M A Mahmoud1,2,
  2. A H S Lee3,
  3. E C Paish3,
  4. R D Macmillan4,
  5. I O Ellis1,3,
  6. A R Green1
  1. 1Division of Pathology, School of Molecular Medical Sciences, University of Nottingham, Nottingham, UK
  2. 2Clinical Pathology Department, Faculty of Medicine, Mansoura University, Mansoura, Egypt
  3. 3Department of Histopathology, Nottingham University Hospitals NHS Trust, Nottingham, UK
  4. 4The Breast Institute, Nottingham University Hospitals NHS Trust, Nottingham, UK
  1. Correspondence to Dr Andrew R Green, Division of Pathology, School of Molecular Medical Sciences, University of Nottingham, Queen's Medical Centre, Nottingham NG7 2UH, UK;{at}


Background Macrophages constitute a major component of the leucocytic infiltrate of tumours. Human studies show an association between tumour-associated macrophages and tumours with poor prognostic features. In breast cancer, the presence of macrophages has been correlated with increased angiogenesis and poor prognosis but little information is available about the independent prognostic role of macrophages infiltrating breast carcinomas.

Aims and methods This study used immunohistochemistry and tissue microarrays to assess the density and localisation of CD68 macrophages infiltrating 1322 breast tumours and to identify any relationship with clinicopathological factors and patient outcome.

Results Tumour-infiltrating macrophages were present in the majority of tumours with a predominantly diffuse pattern. The density of distant stromal macrophages (infiltrating stroma away from the carcinoma, median count 14 cells) was higher than intratumoural (median zero cells) and adjacent stromal macrophages (median three cells). Higher total macrophage number was associated with higher tumour grade (rs=0.39, p<0.001), ER and PgR negativity, HER-2 positivity and basal phenotype (p<0.001). In univariate survival analysis, higher numbers of CD68 macrophages were significantly associated with worse breast cancer-specific survival (p<0.001) and shorter disease-free interval (p=0.004). However in multivariate model analysis, the CD68 macrophage count was not an independent prognostic marker.

Conclusions Macrophages are heterogeneous with different subsets having different functions. The present study suggests that overall macrophage numbers are not related to prognosis in breast cancer. However, further studies are needed to investigate the potential role of different subsets of macrophages.

  • Angiogenesis
  • breast
  • breast cancer
  • breast pathology
  • cancer
  • CD68
  • cytology
  • diagnosis
  • endocrinology
  • histopathology
  • immunocytochemistry
  • immunohistochemistry
  • immunology
  • molecular biology
  • molecular oncology
  • molecular pathology
  • panmacrophage
  • PCR
  • steroid receptors
  • tamoxifen
  • tumour angiogenesis
  • tumour-associated macrophages
View Full Text

Statistics from

Macrophages constitute a heterogeneous population that originates from circulating monocytes and plays a broad role in the maintenance of tissue homeostasis, through the clearance of senescent cells and the remodelling and repair of tissues after inflammation.1 2 Macrophages are a major component of the leucocytic infiltrate of tumours. They can contribute to tumour destruction or facilitate tumour growth and metastasis, depending on their phenotype.3 M1 macrophages are proinflammatory and are characterised by the production of nitric oxide and type 1 cytokines and chemokines,4 while M2 macrophages release anti-inflammatory molecules such as interleukin (IL) 4, IL-13, and/or transforming growth factor β and support angiogenesis.5 6 Mantovani and colleagues7 8 have suggested that tumour-associated macrophages (TAM) are biased towards the M2 type and show mostly pro-tumour functions, promoting tumour cell survival, proliferation and dissemination.

In breast cancers, macrophages are found in abundance and their presence has been correlated with increased angiogenesis and poor prognosis.9–12 Furthermore, genes associated with macrophage infiltration are part of molecular signatures that herald poor prognosis in node-negative, tamoxifen-treated breast carcinomas.13 However, only one study showed an independent role of macrophage prognostic value in multivariate analysis using relatively small patient numbers (91),9 while a more recent larger study did not show this association.14

As macrophages have diverse functions, including being a part of the cellular immune response against tumours and the promotion of tumour progression, it was difficult to propose one hypothesis regarding their role. They could be good or bad prognostic markers. Therefore, the aim of this study was to investigate breast cancer infiltration by macrophages in terms of count, pattern and micro-anatomical localisation and to determine their prognostic value and correlation with various clinicopathological variables in a large cohort of breast cancer patients with long-term follow-up.

Patients and methods


This retrospective study was based on a consecutive series of 1902 cases of primary invasive breast carcinoma diagnosed between 1987 and 1998 and entered into the Nottingham Tenovus Primary Breast Carcinoma series. This is a well-characterised series of patients under the age of 71 years with a long-term follow-up, which has been previously investigated for a wide range of protein expression15 16 and infiltrating lymphocytes.17 18

Patient management was based on tumour characteristics assessed by using the Nottingham prognostic index and hormone receptor status as previously described.17 Survival data were maintained prospectively. Breast cancer-specific survival (BCSS) was defined as the time from the date of primary surgical treatment to the time of death from breast cancer. Disease-free interval (DFI) was calculated from the date of first operation until the first recurrence (local, regional, or distant). This work was approved by Nottingham Research Ethics Committee 2 under the title of ‘Development of a molecular genetic classification of breast cancer’. The reporting recommendations for tumour marker prognostic studies (REMARK) criteria, recommended by McShane et al19 were followed throughout this study.

Tissue microarrays and immunohistochemistry

Breast cancer tissue microarrays (TMA) were constructed from formalin-fixed, paraffin-embedded tumour samples. Cores, 0.6 mm in diameter, were precisely arrayed into a new recipient paraffin block using the manual tissue arrayer from Beecher Instruments (Sun Prairie, Wisconsin, USA), as described previously.20 Freshly cut 4-μm thick breast cancer TMA and full-face sections were immunostained using the standard streptavidin–biotin complex method as described elsewhere.17 Macrophages in 30 full-face breast cancer tissue sections (randomly chosen from those represented on the TMA) and TMA were labelled using an optimised monoclonal anti-CD68 antibody (a pan-macrophage marker,21 clone KP1; Dako, Glostrup, Denmark; dilution 1:3000). Antigen retrieval was performed by microwave treatment in 10 mmol/l sodium citrate buffer (pH 6.0) for 20 min. Paraffin sections of normal human tonsil were used as a positive control. For the negative controls, the primary antibody was omitted.

Macrophage quantification method

The number of CD68 cells was counted in each tumour core using an eyepiece graticule without any previous knowledge of the patients' clinical data (SMAM). CD68 cells were counted in three locations in each tumour; intratumoural, adjacent stroma and distant stromal compartments (defined elsewhere,17 18 figure 1A). The total number of CD68 cells was determined by adding up the counts for the three compartments. Scores were randomly re-examined by the same investigator (SMAM) after a period of time to ensure reproducibility. Scores were also rechecked randomly by a second observer (IOE). Excellent intra-observer and interobserver agreement was found (κ=0.9).

Figure 1

CD68 macrophages infiltrating breast tumour tissue microarray cores. (A) CD68 macrophages intratumoural (black arrow), in adjacent (chevron arrow), and in distant (arrow heads) stromal compartments (×400). (B) Dense macrophage infiltration (×200). (C) Sparse CD68 macrophage infiltration (×200).

Statistical analyses

Assessment of the correlation between the number of CD68 macrophages and clinicopathological parameters was performed using Spearman rank order correlation or the Mann–Whitney U test when appropriate. Survival probabilities were studied by the Kaplan–Meier method, and differences in survival time between patient subgroups were analysed with a log-rank test. Multivariate survival analysis was performed using a Cox proportionate hazard model.22 The optimal cut-off point for CD68 cell count against BCSS and DFI in each tumour compartment was defined using X-tile bio-informatics software (version 3.4.7; Robert L Camp, 2005)23 To avoid the problem of multiple cut-point selection, the software performs standard Monte Carlo cross-validation to produce corrected p values.23 24

All statistical analyses were conducted using the software package SPSS for Windows, V.15. All statistical analyses were two-sided and significance was defined as p<0.05. Cut-off values for ERα, PgR, and HER2 included in this study were chosen before statistical analyses as previously described for this patient series.25 26 Basal phenotype positivity was defined as the detection of expression in 10% or more of invasive malignant cells for CK5/6 and/or CK14.27


CD68 immunostaining in breast cancer

The evaluation of full-face tissue sections revealed macrophages observed in both tumour nests and stroma of breast cancer with a predominantly diffuse pattern. Tumour-infiltrating macrophages were large irregular cells with oval to round nuclei and fine processes that showed strong cytoplasmic staining (sometimes granular) but no nuclear staining for CD68. No other cell types were stained. The stained TMA exhibited similar immunoreactivity in terms of character, distribution and pattern of infiltration (figure 1). Some tumours were extensively infiltrated by macrophages (figure 1B), whereas others showed only sparse macrophage infiltration (figure 1C).

A total of 1322 tumours was available for immunohistochemical scoring after the exclusion of uninformative cores. Fifty-nine per cent of the scored tumours were of no special type; the remaining tumours were tubular mixed (17%), medullary (3%), or other types (21%). Fifty one per cent of the tumours were pathologically high grade (grade 3), whereas 32% and 17% were of grades 2 and 1, respectively. All but two tumours contained CD68 cells. The infiltration pattern was mainly diffuse in the three compartments (96% of intratumoural and 97% of adjacent and distant stromal CD68 cells). The density of distant stromal macrophages (median CD68 count 14 cells) was higher than intratumoural (median zero cells) and adjacent stromal macrophages (median three cells; table 1).

Table 1

CD68 macrophages infiltrating breast tumour micro-anatomy (number of patients 1322)

Correlations between macrophage density and clinicopathological features

A higher total macrophage number was correlated with higher tumour grade and associated with ER and PgR negativity, HER-2 positivity and basal phenotype subgroup (table 2). The total macrophage count was also significantly correlated with larger tumour size and weakly correlated with vascular invasion (table 2). Macrophage counts in the three tumour compartments retained the significant association with grade, ER status and basal phenotype (data not shown).

Table 2

Correlation between total CD68 macrophage count and clinicopathological variables in patients with invasive breast carcinoma

TAM and patient disease-specific survival

The optimal cut-off point for total number, distant stromal and adjacent stromal CD68 cells, determined by X-tile, was 17, whereas that for intratumoural CD68 cells was six. The median follow-up period for the study patients was 127 months; during which 372 patients died from breast cancer and 536 had a recurrence.

In univariate survival analysis, worse BCSS was associated with higher numbers of total, distant and adjacent stromal CD68 cells but not with intratumoural CD68 cells (figure 2). Similar results were found for the associations between CD68 cells in the different locations and DFI in univariate analysis (data not shown).

Figure 2

Kaplan–Meier estimators showing breast cancer-specific survival (BCSS), stratified by the defined cut-off points, for macrophages infiltrating breast carcinomas. (A) Total number, (B) intratumoural, (C) distant stromal and (D) adjacent stromal CD68 macrophages.

In multivariate analysis, higher total macrophage counts were not associated with BCSS or DFI (table 3), after adjustment for tumour size, grade, nodal stage and vascular invasion. Similarly, macrophage counts in the three tumour microenvironments were not associated with survival in the same multivariate model. Of note, when the total number of macrophages was assessed in multivariate analysis in combination with tumour grade, the macrophage count was no longer associated with survival.

Table 3

Multivariate Cox proportional hazard model for effects of the total number of CD68 macrophage infiltrates, in addition to the established prognostic factors, on BCSS and DFI in breast carcinomas


In agreement with the previously published data in human tumours including breast cancer, macrophages infiltrated almost all of the tumours in the present study population. Monocytes are recruited to the tumour microenvironment in breast cancer by molecules produced by neoplastic and/or stromal cells. The chemokine CC-motive ligand 2 (CCL2 or MCP-1), produced by tumour cells and macrophages, is the main protagonist.28 29 When monocytes (then macrophages) reach the tumour mass their differentiation is influenced by several signals from neoplastic cells and the surrounding stroma for example, IL-10 and transforming growth factor β. In the present study, macrophages were found to infiltrate all tumour compartments (intratumoural, adjacent and distant stroma) with higher densities in the distant stroma. Higher macrophage counts in the stroma may be due to higher concentrations of IL-10.30

Although the KP1 clone of anti CD68 antibody was reported to stain some granulocyte precursors,31 it has been used to identify macrophages in routinely fixed paraffin-embedded tissues from many cancer types including breast cancer.32–34 This is the first study to assess macrophage infiltration in TMA from breast tumours, and therefore the scoring methodology was different from that previously adopted by Leek et al9 and Lee et al35 to score full-face tissue sections. In their studies, there was no comment on the micro-anatomical localisation of macrophages, but in agreement with the current results they reported a diffuse pattern of infiltration.

The density of CD68 macrophages was correlated with higher tumour histological grade in agreement with previous studies.11 35 36 Multivariate analysis is thus essential when examining the relation of macrophage infiltration to survival. In agreement with Tsutsui et al,37 macrophage density was associated with a poor prognosis in univariate but not in multivariate analysis. This difference is a reflection of the association between the macrophage count and grade. Our results were in contrast to the findings of Leek et al,9 who detected an independent poor prognostic role for macrophages in breast cancer. However, they used a much smaller study of only 101 patients. Genetic analysis of TAM has recently provided evidence that an enrichment of macrophage transcripts is predictive of poor prognosis and reduced survival in human breast cancer.38 Macrophages have the most complex transcriptome known.39 Different macrophage subpopulations may have different pro or antitumour functions. CD68 is a panmacrophage marker that stains macrophages in different situations and probably of different function.40 Assessment of the monocyte/macrophage-restricted membrane proteins CD163 and/or CD204 in addition to CD68 would probably be informative as the expression of both has been associated with the anti-inflammatory M2 macrophage phenotype.41 Furthermore, macrophages need to be activated before they can induce their effects.42 Therefore, an assessment of macrophage function, rather than phenotype, may be more meaningful in predicting patient outcomes. Thymidine phosphorylase in macrophages is associated with a worse prognosis in breast cancer.43 44 In a recently published study, Campbell et al45 found that proliferating macrophages did not have an independent prognostic significance in each of two separate cohorts although both were small (110 and 106 patients).

In a recent paper, DeNardo et al46 found an inverse correlation between the number of CD68 and CD8 cells in breast cancer. They also found a better prognosis for patients with low CD68 and high CD8 counts in their carcinoma and a worse prognosis for patients with high CD68 and low CD8 counts, including the subset of patients who received chemotherapy. By contrast, we found a positive correlation between CD68 and CD8 numbers in our dataset (rs=0.46, p<0.001) and in another dataset (rs=0.55, p<0.001 (unpublished analysis of Lee et al).35 We have previously found that CD8 is an independent prognostic factor.18 We therefore investigated whether the CD68 count had an additional effect. In multivariate analysis including traditional prognostic factors and CD8, CD68 was not an independent prognostic factor in the whole patient group or the subset who received chemotherapy. In addition, CD68 had no prognostic impact in subsets with low or high CD8 (data not shown).

The results of this study suggest that tumour-infiltrating CD68 macrophages are not an independent prognostic factor in this breast cancer series perhaps due to the association with higher grade. Assessment of CD68 macrophages in different areas of the tumour micro-anatomy also did not have a significant association with patient outcomes. The present study provides, for the first time, robust information about TAM in a large series of breast carcinomas, but future studies of macrophage subpopulations are warranted.

Take-home messages

  • Tumour-infiltrating CD68 macrophages are not an independent prognostic factor in this breast cancer series perhaps due to the association with higher grade.

  • The present study provides, for the first time, robust information about TAM in a large series of breast carcinomas, but future studies of macrophage subpopulations are warranted.


View Abstract


  • Competing interests None.

  • Ethics approval Ethics approval was provided by Nottingham Research Ethics Committee 2 under the title of development of a molecular genetic classification of breast cancer.

  • Provenance and peer review Not commissioned; externally peer reviewed.

Request Permissions

If you wish to reuse any or all of this article please use the link below which will take you to the Copyright Clearance Center’s RightsLink service. You will be able to get a quick price and instant permission to reuse the content in many different ways.