Targeting Tumour-Associated Macrophages/Microglia

In a paper by HF Dvorak published in 1986, tumors were memorably described as “wounds that do not heal”. A re-phrasing of this would perhaps be even more accurate: cancer behaves like a wound which never stops healing, and co-opts the body’s natural healing abilities for pathological growth and expansion. This is the nature of the cancer-inflammation connection.

Contrary to what one might expect, certain cells of the immune system, actively modified by the tumour, often do more harm than good in the case of gliomas. The immune system is a complex community composed of many different cell types with a wide variety of functions. Tumours have the unfortunate ability ability to re-train immune cells (such as macrophages) from a cancer-fighting to a cancer-promoting role.

Macrophages (meaning “big eaters” in Greek) are immune cells which phagocytose (or eat) cellular debris and pathogens. The primary immune cells found in a glial tumour are the tumour-associated macrophages and microglia (TAMs). Whereas the microglia are the resident macrophages of the central nervous system, non-resident circulating macrophages are also found within tumours.

Cytokines are a broad class of intercellular signaling molecules, or chemical messengers. The macrophages and microglia are actively recruited into the tumour, following the trail of chemoattractant cytokines (chemokines) secreted by the tumour cells.

Glioma cells further alter the behaviour of immune cells by secreting more cytokines which reprogram these macrophages and microglia to an alternative cancer-promoting phenotype (in this case called M2). Cancer cells have an uncanny ability to co-opt and utilize natural cellular behaviours to further their own expansive growth. In this case, glioma cells cause the TAMs to adopt a phenotype close to that observed in the process of wound healing.

As stated in a 2005 study describing the overlap in the genes upregulated in wound healing and cancer, “both wound repair and cancer are characterized by cell proliferation, remodeling of extracellular matrix, cell invasion and migration, new blood vessel formation, and modulation of blood coagulation.” (1)

Through the secretion of cytokines, many of which are also secreted by the tumour, these alternatively activated M2 microglia and macrophages, conditioned by the tumour cells, actively promote tumour invasiveness, angiogenesis, and the suppression of the cancer-fighting abilities of other immune cells such as T cells (2, 3, 4).

The transcription factor STAT3, previously discussed in the Targeting Invasion and Targeting Glioma Stem Cells sections, is also a crucial factor in the orchestration of the pro-cancer immune response. As one author phrased it, “the distinction between immune responses that inhibit and those that promote cancer is, to a large extent, regulated by members of the STAT family of proteins.” (5)

“Continuous expression of wound healing genes by persistently activated STAT3 in tumor cells may be essential for malignant progression.” (1)

STAT3 is not only upregulated in glioma cells but also in glioma-associated immune cells such as TAMs. Therefore, inhibition of STAT3 is also a very important way to combat the immunosuppression associated with cancer. Additional pharmacological approaches will be discussed below.

Who is most likely to benefit from this strategy

Before discussing drugs which target cancer-promoting activites of TAMs, we must first explore exactly who is most likely to benefit, and at what time during tumour progression.

A fascinating study published in 2013 analyzed gene transcripts of hundreds of glioblastoma samples from the Cancer Genome Atlas (6). Specifically, the researchers were looking for relative expression of immunity-related genes and how these might differ between different glioblastoma subtypes. They found that the greatest degree of immunological variation was found between the mesenchymal and proneural glioblastoma subtypes. The mesenchymal subtype, known as the most highly angiogenic type, was also found to be the most immunologically reactive, with a relative overexpression of both immunosuppressive and immune effector genes. The authors proposed that this observed increase in immunosuppression in the mesenchymal subtype may be a tumour response to the already increased immune effector activity in these tumours. This subtype may be the most responsive to certain immunotherapies, as observed in one small trial of dendritic cell therapy (7).

Notably, mesenchymal-type tumours were enriched in gene expression associated with macrophage and monocyte (macrophage precursor) activity, including monocyte attracting chemokines (CCL2 and CCL22), markers of immunosuppressive M2 macrophages (CD 163), and arginase (an inhibitor of T cell responses secreted by myeloid derived suppressor cells). Additionally, expression of genes involved in the STAT3 pathway were also enriched in the mesenchymal subtype relative to other subtypes (6).

Another study with mice showed that the mesenchymal transition, which is frequently found in recurrent human glioblastomas, is associated with increased macrophage infiltration into the tumour, increased angiogenesis and STAT3 activity (8).

Further evidence that the degree of immune infiltration into a tumour is subtype dependent was provided in a study (12) which analyzed 19 human glioblastoma samples for markers of mesenchymal and proneural subtypes. The categorized tumours (10 mesenchymal and 9 proneural) were then analyzed for infiltration of cytotoxic T-lymphocytes and microglia. The average number of both of these types of immune cells was significantly increased in the mesenchymal-type tumours. This study also demonstrated that proneural-type tumours have an increased transforming growth factor-beta (TGFb) mediated immunosuppression, perhaps explaining the failure of this subtype to respond to dendritic cell immunotherapy in one small trial (7).

An additional study attempted to understand the dynamics of immune infiltration and function in an immunocompetent mouse model of glioblastoma (9). First, the relative frequencies of tumour-infiltrating lymphocytes (including T helper and cytotoxic T lymphocytes) versus tumour-associated macrophages/microglia (TAMs) was assessed at 10 days following tumour implantation. TAMs were found at much higher frequencies than infiltrating lymphocytes in each of the animals. Further investigations, at days 13, 26 and 40 showed the time-dependent dynamics of immune infiltration and activity. Very small changes in infiltrating lymphocyte frequency, and TAM frequency and function were observed between the early and intermediate time points. In contrast, at the late time point (40 days after implantation), marked increases in tumour-infiltrating lymphocytes and TAMs were seen. At this late stage, there was also a marked decrease in the secretion of tumour necrosis factor-alpha (TNF-a) by the macrophages, indicating a transition from the tumour-fighting M1 to the tumour-promoting M2 phenotype. There was an early increase in the immunosuppressive Treg population in the spleen, and a corresponding increase in the immunosuppressive cytokine interleukin-10 (IL-10). These changes occurred far in advance of the local increase in TAMs within the tumour itself.

In summary these studies demonstrate that the immunoreactive mesenchymal subtype of tumour (most common in glioblastoma) will likely derive the most benefit from a TAM-targeting strategy, though other subtypes may benefit as well. Tumours of other subtypes may transition to a mesenchymal-type pattern at the time of recurrence, which brings with it an influx of cancer-promoting M2 macrophages and microglia into the tumour microenvironment.

Targeting TAMs with repurposed drugs


Minocycline is a tetracycline antibiotic which has been approved for decades to treat conditions such as acne and rosacea.

One of the ways that microglial cells (the resident macrophages of the central nervous system) may aid in tumour invasion is through the secretion of membrane type-1 matrix metalloprotease (MT1-MMP). MT1-MMP cleaves the inactive form of MMP-2 secreted by glioma cells, converting into an active form. MMP-2 is an extracellular matrix degrading enzyme, necessary for tumour invasion. Another matrix-degrading enzyme, MMP9, may also be produced by tumour-associated microglia. Minocycline has been shown in two studies (10, 13) to inhibit the production of pro-invasive enzymes by tumour-associated microglia.

In a high quality animal model of glioblastoma using the GL261 cell line, test animals were given oral minocycline hydrochloride in their drinking water (10). In the test mice, tumour volume was reduced by 78% compared to control mice when minocycline was given immediately after tumour cell implantation and continued for two weeks. When tumours were allowed to develop for one week, at which time minocycline was administered for two weeks, tumour volumes were reduced by 46%, indicating an advantage to an earlier administration of the drug.

Propentofylline (Vivitonin)

Propentofylline is a synthetic xanthine derivative marketed as Vivitonin, which is used to treat lethargy in dogs. It increases blood flow to and oxygenation of the brain. In human studies it has shown minimal adverse effects.

A highly interesting study using the CNS-1 rat glioblastoma model showed that propentofylline significantly decreased tumour growth through effects on microglia, rather than direct effects on glioma cells (11). Rats in the study were implanted with CNS-1 cells and administered propentofylline for 14 days beginning on the day of tumour implantation. On day 14 average tumour volume was reduced by 88% compared to the control group, while over 60% of the treated rats didn’t develop a detectable tumour at all. Tumours in the treated animals also remained as intact entities, without evidence of diffusion, which is remarkable for this highly invasive model. Survival was significantly prolonged in the treated animals. Rats were also implanted with CNS-1 and the tumours were allowed to grow until detectable by MRI before administration of propentofylline. According to Figure 8 of the published study, tumour volumes did not increase upon administration of propentofylline, while control tumours continued to grow rapidly.

In this study, the beneficial effect of propentofylline was attributed to its prevention of microglial migration towards the tumour, and the blocking of MMP-9 (important for tumour invasion) expression by microglia. In vitro, propentofylline had no effect on tumour cell invasion when cultured without microglia, while invasion was significantly inhibited by the drug when tumour cells were co-cultured with microglia, demonstrating the microglia-specific effects of the drug.

The CNS-1 model was chosen for this study due to its highly proliferative and invasive nature, and for its similarity to human glioblastoma, despite being a rat glioma. It is a model which can be used in immunocompetent rats, and is therefore excellent for testing immunotherapies in glioblastoma. The results of this study are most applicable to highly invasive mesenchymal glioblastomas. The possible effects of the drug for lower-grade gliomas remains to be tested.


Taken together, these studies show that in highly invasive gliomas, the therapeutic targeting of tumour-associated macrophages and microglia is crucial. Glioblastomas, particularly of the mesenchymal subtype, are most likely to benefit. Recurrent glioblastomas and possibly lower grade tumours tend to undergo a mesenchymal transition, evident at the time of recurrence. It is not yet clear to what degree TAMs contribute to the progression of lower grade gliomas, though it is likely that there would be some benefit to strategies targeting TAMs, especially in recurrent cases or cases where increased invasion or angiogenesis is evident.

  1. Stat3 regulates genes common to both wound healing and cancer. Dauer et al. 2005.
  2. Tumor-associated macrophages in glioma: friend or foe? Kennedy et al. 2013.
  3. Microglia and macrophages in malignant gliomas: recent discoveries and implications for promising therapies. Carvalho da Fonseca et al. 2013.
  4. The controversial role of microglia in malignant gliomas. Wei et al. 2013.
  5. STATs in cancer inflammation and immunity: a leading role for STAT3. Yu et al. 2009.
  6. Immune heterogeneity of glioblastoma subtypes: extrapolation from the cancer genome atlas. Doucette et al. 2013.
  7. Gene expression profile correlates with T-cell infiltration and relative survival in glioblastoma patients vaccinated with dendritic cell immunotherapy. Prins et al. 2011.
  8. Signal transducer and activator of transcription 3 promotes angiogenesis and drives malignant progression in glioma. Doucette et al. 2012.
  9. Dynamics of central and peripheral immunomodulation in a murine glioma model. Kennedy et al. 2009.
  10. Minocycline reduces glioma expansion and invasion by attenuating microglial MT1-MMP expression. Markovic et al. 2011.
    READ ABSTRACT Email me for a PDF copy
  11. Propentofylline decreases tumor growth in a rodent model of glioblastoma multiforme by a direct mechanism on microglia. Jacobs et al. 2011.
  12. The cancer stem cell subtype determines immune infiltration of glioblastoma. Beier et al. 2012.
  13. Glioma-associated microglial MMP9 expression is upregulated by TLR2 signaling and sensitive to minocycline. Hu et al. 2014.
    READ ABSTRACT Email me for a PDF copy