Part V · Immunology and the tumor microenvironment · Chapter 23

Functional states of the microenvironment

Why releasing a checkpoint does so little here, stated as mechanism rather than as disappointment.

1 · Inflammation and local immunosuppression

Inflammation in the tumour microenvironment is not a single state and the word obscures more than it conveys. Acute inflammation with antigen presentation and effector recruitment is what immunotherapy tries to produce. Chronic low-grade inflammation with myeloid recruitment and tissue remodelling is what tumours produce, and the two share vocabulary while working in opposite directions.

The suppressive programme has identifiable components. IL-6 and related cytokines sustain myeloid recruitment and a chronic wound-healing phenotype. TGF-beta is the most consequential single mediator, because it suppresses effector function, promotes regulatory T-cell differentiation and drives the matrix deposition that physically excludes T cells. In a large analysis of checkpoint blockade, TGF-beta signalling in fibroblasts was associated with T-cell exclusion and with failure of PD-L1 blockade¹.

The clinical correlate in breast cancer is inflammatory breast cancer, which is a distinct entity with its own dismal prognosis² and is discussed in Locally advanced and inflammatory breast cancer. The name describes the clinical presentation rather than the immunological state, and conflating the two is a common error.

Systemic inflammation contributes separately. Neutrophil extracellular traps generated during systemic inflammatory events can reactivate dormant cells³, which means an inflammatory episode remote from the tumour is a plausible mechanism for the late recurrences in Dormancy, residual disease, and late recurrence.

Interferon signalling occupies an awkward position. Acute interferon response supports antigen presentation and effector function. Chronic interferon signalling drives adaptive resistance including PD-L1 upregulation, and diminished interferon signalling with age associates with checkpoint blockade outcomes⁴.

2 · Hypoxia, acidosis, and nutrient competition

Three separate constraints are usually collapsed into one word. They have different evidence and different consequences, so they are worth separating.

Hypoxia comes first. A randomised trial of primary epirubicin with or without tamoxifen enrolled 187 patients with T2 to T4 breast cancer. Overall response fell progressively as tumour HIF-1α expression by immunohistochemistry rose, and HIF-1α was an independent predictor of response⁵. It was also associated with significantly shorter disease-free survival. That survival association held in oestrogen-receptor-positive patients and not in receptor-negative patients, which is the luminal deficit appearing in an unexpected place.

Acidosis is a separate quantity from oxygen tension and should be measured as one. Intravital monitoring of peritumoural pH found that the regions of highest invasion corresponded to the regions of lowest pH, and that invasion did not occur into tissue with near-normal extracellular pH⁶. Oral sodium bicarbonate raised peritumoural pH and inhibited local invasion in a preclinical model.

Nutrient competition is the third, and it connects most directly to immunology. In a mouse sarcoma model, glucose consumption by tumour cells metabolically restricted T cells, lowering their mTOR activity, glycolytic capacity and interferon gamma production⁷. Antibodies against CTLA-4, PD-1 and PD-L1 restored glucose in the tumour microenvironment and permitted T cell glycolysis.

That last result reframes what a checkpoint antibody does. Part of its effect is metabolic rather than purely a matter of receptor signalling, and part of it is exerted on the tumour cell rather than on the T cell.

3 · Checkpoint biology beyond the PD-1 and PD-L1 axis

PD-1 and PD-L1 are one pair of molecules in a system with many. CTLA-4 acts earlier in the response and largely in the lymph node rather than in the tumour. LAG-3, TIM-3 and TIGIT are co-inhibitory receptors with their own ligands and their own expression kinetics⁸.

The adenosine axis has the best breast-specific evidence. CD73 converts extracellular AMP to adenosine, which suppresses T cell function through the A2A receptor. Using gene expression data from more than 6,000 patients with breast cancer, high CD73 expression was associated with poor prognosis in triple-negative disease⁹. Among patients with triple-negative disease given anthracycline-only preoperative chemotherapy, high CD73 expression was associated with a lower rate of pathological complete response. In mouse models, CD73 overexpression conferred resistance to doxorubicin by suppressing adaptive antitumour responses through the A2A receptor, and CD73 blockade restored doxorubicin-mediated antitumour immunity.

That is an immunological mechanism of chemotherapy resistance. It is worth noticing, because cytotoxic resistance is rarely discussed in these terms.

One measurement point applies to every molecule in this section. Expression of a checkpoint is not a measurement of its activity. PD-L1 assays differ in antibody, in which cells are scored and in threshold, so two tumours called positive by different assays are not interchangeable. The analytic validity framework is in Receptor assessment and the measurement problem, and the clinical use of these agents is in Immunotherapy.

4 · Hormone receptor driven immune evasion and antigen presentation

The luminal deficit is usually described. This section asks whether it can be explained.

The description is settled. Hormone-receptor-positive HER2-negative tumours have the lowest proportion of lymphocyte-predominant cases, the lowest CD8 infiltration, and the lowest tumour mutation burden of the common subtypes^10,11. The prognostic direction of the infiltrate also reverses, with each 10% increment in lymphocytes carrying an overall survival hazard ratio of 1.10 in luminal HER2-negative disease against 0.92 in triple-negative disease¹².

Low antigenicity accounts for part of it. A tumour with few mutations presents few neoantigens, and nothing downstream can compensate for an absent target.

It does not account for all of it. The strongest evidence that oestrogen signalling is an active immune-evasion programme comes from models in which the tumour itself is oestrogen-insensitive. In those models, oestrogen signalling accelerated tumour progression by mobilising myeloid-derived suppressor cells and by enhancing their intrinsic suppressive activity¹³. The effect required an intact immune system. It disappeared in immunodeficient hosts and on depletion of those cells. Mechanistically, oestrogen receptor alpha activated STAT3 in bone marrow myeloid precursors by enhancing JAK2 and SRC activity.

Read carefully, that is a claim about the host rather than about the tumour. Oestrogen deprivation may therefore have an immunological effect independent of its effect on the tumour cell, and an aromatase inhibitor alters the host myeloid environment as well as the ligand supply. The receptor biology those drugs are designed around is in Estrogen receptor signaling and endocrine resistance.

Tumour-intrinsic suppression of antigen presentation by oestrogen receptor signalling is frequently proposed as the remaining explanation. It is a plausible mechanism and it is considerably less well established than the myeloid one. It belongs in this chapter as a hypothesis, because the two readings predict different interventions.

5 · HER2-driven evasion, CD47 and p95HER2

HER2-positive disease presents two evasion problems specific to it, and both concern the effector arm rather than the T cell.

The first is the truncated receptor. p95HER2 is an amino-terminally truncated form that retains kinase activity and lacks the extracellular region trastuzumab binds. Among 46 patients with metastatic breast cancer treated with trastuzumab, 1 of 9 whose tumours expressed p95HER2 responded, against 19 of 37 whose tumours expressed the full-length receptor¹⁴. Lapatinib inhibited p95HER2 phosphorylation and growth in models where trastuzumab had no effect.

The immunological reading is the one usually missed. An antibody that cannot bind cannot present its Fc region to an effector cell. A tumour expressing p95HER2 is antigen negative for the purposes of antibody-dependent cytotoxicity while remaining HER2 amplified by any nucleic acid assay. That is the compartment problem of HER2 heterogeneity arriving through a different door, and the receptor biology is in HER family signaling.

The second is the macrophage checkpoint. CD47 on the tumour cell engages SIRPα on the macrophage and inhibits phagocytosis. Combining CD47 blockade with trastuzumab suppressed the growth of HER2-positive breast cancer models that had become tolerant to antibody-dependent cellular cytotoxicity, and the effect was Fc dependent and mediated by phagocytosis¹⁵.

Two limits belong with that. The work is preclinical. And CD47 is expressed on normal cells including erythrocytes, so the open question is therapeutic index rather than mechanism.

6 · TGF-beta and metabolic immunosuppression

TGF-β is the clearest single link between the stromal architecture of Cellular architecture of the tumor microenvironment and the functional states described here.

The pivotal clinical observation is not from breast cancer. In metastatic urothelial cancer treated with atezolizumab, lack of response was associated with a signature of TGF-β signalling in fibroblasts¹. In those tumours CD8 T cells sat in collagen-rich peritumoural stroma rather than in tumour parenchyma. In a mouse model reproducing that phenotype, blocking TGF-β together with PD-L1 allowed T cells into the tumour centre and produced regression.

Breast cancer contributes the cellular description rather than the intervention. Among eight CAF-S1 clusters identified in breast tumours, the cluster defined by TGF-β signalling was one of two indicative of primary resistance to immunotherapy¹⁶.

Metabolic suppression runs alongside it by separate means. Glucose depletion restricts T cell effector function⁷. Adenosine generated by CD73 suppresses it through a receptor⁹.

The common feature is the one to carry forward. None of these mechanisms is reversed by blocking PD-1. A T cell starved of glucose, held in adenosine, or physically excluded by a TGF-β-driven matrix is not a T cell whose inhibitory receptor is the limiting problem. That is the mechanistic form of the argument in Primary and acquired resistance to checkpoint blockade.

7 · MHC loss and antigen presentation defects

An immune response against a tumour requires that the tumour present antigen. Loss of that capability is the most direct escape mechanism available and one of the most common.

The pathway has several separable failure points and they are not equivalent. Loss of beta-2 microglobulin abolishes surface expression of all class I molecules, is usually irreversible because it is genetic, and cannot be restored by cytokine. Loss of a single HLA allele removes presentation of the peptides restricted to that allele while leaving others intact, which is a partial and antigen-specific escape. Downregulation of the antigen processing machinery, including the transporter associated with antigen processing and the immunoproteasome subunits, reduces presentation transcriptionally and can be reversed by interferon.

The distinction matters therapeutically and is rarely made in clinical reports. A tumour with transcriptional downregulation may be recoverable with interferon or with agents that increase presentation. A tumour with biallelic beta-2 microglobulin loss is not, and no T-cell-directed therapy will work against it. That includes checkpoint blockade, T-cell engagers and CAR-T, which all converge on a T cell that requires a target it can see.

This is why natural killer cell-directed strategies are mechanistically interesting rather than merely alternative. Natural killer cells recognise the absence of class I, so the escape mechanism that defeats T cells is the signal that activates them.

Class II expression on tumour cells is a separate question and is associated in breast cancer with greater immune infiltration and better outcome, consistent with a role in sustaining CD4 help rather than in direct killing.

8 · Vascular permeability, interstitial pressure, and drug delivery

A drug that does not reach a cell has no pharmacology at that cell. This is obvious and it is routinely left out of discussions of resistance.

Solid tumours show raised interstitial fluid pressure, and the contributing factors are separable: abnormal vessels that leak plasma protein, absent or non-functional lymphatics, fibrosis, and contraction of the interstitial matrix¹⁷. The result is a barrier to transcapillary transport, because convective transport needs a pressure gradient and the gradient has been flattened.

Three consequences follow for breast cancer.

Large molecules are affected more than small ones. A monoclonal antibody or an antibody-drug conjugate depends on convection more than a small-molecule inhibitor does, so delivery is a sharper constraint for the classes in Antibody-drug conjugates and targeted delivery.

Stromal density and interstitial pressure are related, so a desmoplastic tumour has a delivery problem and an immune exclusion problem at once. The matrix described in Cellular architecture of the tumor microenvironment produces both.

And the two kinds of resistance that follow look identical in clinic. A tumour that does not respond may contain cells that are insensitive to the drug, or cells that never received it. No routine assay separates them, and a sensitivity assay run on dissociated cells measures only the first.

9 · Microenvironmental determinants of treatment sensitivity

The microenvironment determines response to therapies that are not immunotherapy, and this is the part of the argument most often omitted.

Delivery comes first. The abnormal vasculature and raised interstitial pressure described in Endothelial cells, pericytes, and angiogenesis limit how much drug reaches tumour cells, and dense matrix limits diffusion once it arrives. A drug can be entirely active against a cell in culture and inactive in a tumour for purely physical reasons, which is one of the translation failures in Experimental models and discovery platforms.

Stromal protection is the second mechanism. Fibroblast-derived factors support tumour cell survival during treatment, and the protection is contact-dependent and soluble. This is a route to resistance that involves no change in the tumour cell at all and will therefore not appear on a resistance biopsy, which is the point made in Overcoming heterogeneity and acquired resistance.

Hypoxia is the third. Hypoxic cells cycle slowly and resist drugs targeting proliferation. Radiation requires oxygen for the fixation of DNA damage, so hypoxic regions are intrinsically radioresistant.

Immune dependence is the fourth and the most recently appreciated. Anthracyclines and some platinum agents induce immunogenic cell death, and their full effect depends on an intact immune response. Trastuzumab's mechanism includes antibody-dependent cellular cytotoxicity, which requires effector cells. A drug classed as cytotoxic may therefore perform differently in an immune-excluded tumour than in an inflamed one.

The consequence for measurement is that the spatial immune phenotypes in Spatial organization and what it adds predict response to conventional therapy and not only to immunotherapy^18,19.

10 · Remodeling of the microenvironment during therapy

The microenvironment measured at diagnosis is not the microenvironment present at the end of treatment.

In 58 patients with breast cancer treated with neoadjuvant chemotherapy, stromal lymphocytes and PD-L1 were quantified by quantitative immunofluorescence on paired pre-treatment and residual disease samples²⁰. There was a trend towards higher lymphocyte counts after chemotherapy. An increase in lymphocytes was associated with longer five-year recurrence-free survival. PD-L1 expression, in both stromal and tumour cells, was significantly lower in the post-treatment samples.

Those two directions should be separated rather than summarised as one immunological effect of chemotherapy. Lymphocytes went up. The ligand for the class of drug most likely to be given next went down.

The measurement consequence is the operative one. A biomarker measured on residual disease is not the same biomarker measured at baseline, and a decision rule validated on one does not transfer to the other. Post-neoadjuvant decisions are made on residual disease, and residual cancer burden quantifies how much of it there is. The immune state of that residuum has not been characterised well enough to select therapy from it.

The cohort is 58 patients, mostly hormone receptor positive, and the findings are associations. They are a reason to measure again rather than a reason to act.

11 · Primary and acquired resistance to checkpoint blockade

Primary resistance means the tumour never responded. Acquired resistance means it responded and then stopped. The mechanisms differ, and conflating them has produced a good deal of confused writing.

Primary resistance is the common case in breast cancer, and most of this chapter has described its components. Few neoantigens¹¹. A sparse infiltrate, sparsest in luminal disease¹⁰. Exclusion behind a TGF-β-driven stroma^1,16. Metabolic restriction⁷. Adenosine⁹. None of those is addressed by blocking PD-1, which is the mechanistic reason single-agent checkpoint blockade does so little in unselected breast cancer⁸.

Acquired resistance has the cleaner mechanistic literature, and it comes from melanoma. Four patients who responded to pembrolizumab and later relapsed had paired baseline and relapse lesions sequenced. Two carried loss-of-function mutations in JAK1 or JAK2 with deletion of the wild-type allele. A third carried a truncating mutation in B2M²¹. The first defect leaves the cell unable to respond to interferon gamma. The second removes surface MHC class I.

The logic transfers even though the disease does not. Both routes break the effector step rather than the checkpoint step. A cell that cannot present antigen, or cannot receive an interferon signal, is invisible whatever is done to PD-1.

Where checkpoint blockade does work in breast cancer, it works with chemotherapy. In KEYNOTE-522, pathological complete response was 64.8% with pembrolizumab plus chemotherapy against 51.2% with chemotherapy alone, among the first 602 patients randomised²². More than half of the control arm reached the same endpoint without it. The development of these agents is in Immunotherapy and their use in triple-negative disease in The biomarker triage, and what is left when it finds nothing.

References

All 22 references

1. Mariathasan S, Turley SJ, Nickles D, et al. TGF-beta attenuates tumour response to PD-L1 blockade by contributing to exclusion of T cells. Nature 2018 554:544-548. PMID 29443960
2. Dawood S, Merajver SD, Viens P, et al. International expert panel on inflammatory breast cancer: consensus statement for standardized diagnosis and treatment. Ann Oncol 2011 22:515-523. PMID 20603440
3. Albrengues J, Shields MA, Ng D, et al. Neutrophil extracellular traps produced during inflammation awaken dormant cancer cells in mice. Science 2018 361:eaao4227. PMID 30262472
4. Sceneay J, Goreczny GJ, Wilson K, et al. Interferon signaling is diminished with age and is associated with immune checkpoint blockade efficacy in triple-negative breast cancer. Cancer Discov 2019 9:1208-1227. PMID 31217296
5. Generali D, Berruti A, Brizzi MP, et al. Hypoxia-inducible factor-1alpha expression predicts a poor response to primary chemoendocrine therapy and disease-free survival in primary human breast cancer. Clin Cancer Res 2006 12:4562-4568. PMID 16899602
6. Estrella V, Chen T, Lloyd M, et al. Acidity generated by the tumor microenvironment drives local invasion. Cancer Res 2013 73:1524-1535. PMID 23288510
7. Chang CH, Qiu J, O'Sullivan D, et al. Metabolic competition in the tumor microenvironment is a driver of cancer progression. Cell 2015 162:1229-1241. PMID 26321679
8. Sharma P, Goswami S, Raychaudhuri D, et al. Immune checkpoint therapy: current perspectives and future directions. Cell 2023 186:1652-1669. PMID 37059068
9. Loi S, Pommey S, Haibe-Kains B, et al. CD73 promotes anthracycline resistance and poor prognosis in triple negative breast cancer. Proc Natl Acad Sci U S A 2013 110:11091-11096. PMID 23776241
10. Stanton SE, Adams S, Disis ML. Variation in the incidence and magnitude of tumor-infiltrating lymphocytes in breast cancer subtypes: a systematic review. JAMA Oncol 2016 2:1354-1360. PMID 27355489
11. Barroso-Sousa R, Jain E, Cohen O, et al. Prevalence and mutational determinants of high tumor mutation burden in breast cancer. Ann Oncol 2020 31:387-394. PMID 32067680
12. Denkert C, von Minckwitz G, Darb-Esfahani S, et al. Tumour-infiltrating lymphocytes and prognosis in different subtypes of breast cancer: a pooled analysis of 3771 patients treated with neoadjuvant therapy. Lancet Oncol 2018 19:40-50. PMID 29233559
13. Svoronos N, Perales-Puchalt A, Allegrezza MJ, et al. Tumor cell-independent estrogen signaling drives disease progression through mobilization of myeloid-derived suppressor cells. Cancer Discov 2017 7:72-85. PMID 27694385
14. Scaltriti M, Rojo F, Ocana A, et al. Expression of p95HER2, a truncated form of the HER2 receptor, and response to anti-HER2 therapies in breast cancer. J Natl Cancer Inst 2007 99:628-638. PMID 17440164
15. Upton R, Banuelos A, Feng D, et al. Combining CD47 blockade with trastuzumab eliminates HER2-positive breast cancer cells and overcomes trastuzumab tolerance. Proc Natl Acad Sci U S A 2021 118:e2026849118. PMID 34257155
16. Kieffer Y, Hocine HR, Gentric G, et al. Single-cell analysis reveals fibroblast clusters linked to immunotherapy resistance in cancer. Cancer Discov 2020 10:1330-1351. PMID 32434947
17. Heldin CH, Rubin K, Pietras K, Ostman A. High interstitial fluid pressure: an obstacle in cancer therapy. Nat Rev Cancer 2004 4:806-813. PMID 15510161
18. Gruosso T, Gigoux M, Manem VSK, et al. Spatially distinct tumor immune microenvironments stratify triple-negative breast cancers. J Clin Invest 2019 129:1785-1800. PMID 30753167
19. Hammerl D, Martens JWM, Timmermans M, et al. Spatial immunophenotypes predict response to anti-PD1 treatment and capture distinct paths of T cell evasion in triple negative breast cancer. Nat Commun 2021 12:5668. PMID 34580291
20. Pelekanou V, Carvajal-Hausdorf DE, Altan M, et al. Effect of neoadjuvant chemotherapy on tumor-infiltrating lymphocytes and PD-L1 expression in breast cancer and its clinical significance. Breast Cancer Res 2017 19:91. PMID 28784153
21. Zaretsky JM, Garcia-Diaz A, Shin DS, et al. Mutations associated with acquired resistance to PD-1 blockade in melanoma. N Engl J Med 2016 375:819-829. PMID 27433843
22. Schmid P, Cortes J, Pusztai L, et al. Pembrolizumab for early triple-negative breast cancer. N Engl J Med 2020 382:810-821. PMID 32101663