Chemoimmunotherapy: reengineering tumor immunity

Abstract

Cancer chemotherapy drugs have long been considered immune suppressive. However, more recent data indicate that some cytotoxic drugs effectively treat cancer in part by facilitating an immune response to the tumor when given at the standard dose and schedule. These drugs induce a form of tumor cell death that is immunologically active, thereby inducing an adaptive immune response specific for the tumor. In addition, cancer chemotherapy drugs can promote tumor immunity through ancillary and largely unappreciated immunologic effects on both the malignant and normal host cells present within the tumor microenvironment. These more subtle immunomodulatory effects are dependent on the drug itself, its dose, and its schedule in relation to an immune-based intervention. The recent approvals of two new immune-based therapies for prostate cancer and melanoma herald a new era in cancer treatment and have led to heightened interest in immunotherapy as a valid approach to cancer treatment. A detailed understanding of the cellular and molecular basis of interactions between chemotherapy drugs and the immune system is essential for devising the optimal strategy for integrating new immune-based therapies into the standard of care for various cancers, resulting in the greatest long-term clinical benefit for cancer patients.

Introduction

Cancer therapy is designed to specifically integrate distinct treatment modalities in the most effective way to achieve the highest cure rate. Surgery and radiation therapies are used for locoregional disease control, whereas systemic therapies are used to treat micrometastatic or widespread metastatic cancers and hematologic malignancies. While systemic therapies have historically been given after local measures have been undertaken to remove the primary tumor, they are increasingly used prior to definitive local treatment both to achieve systemic disease control earlier and to evaluate the responsiveness of the tumor to treatment. Regardless of the timing in relation to local therapy, these systemic treatments—chemotherapy, endocrine therapy, small molecular targeted therapies and monoclonal antibodies–are designed to decrease the likelihood of relapse due to micrometastatic disease.

Therapies that induce an immune response to tumors—immunotherapies—have been investigated for over 100 years as attractive strategies for cancer treatment [1]. While initially tested alone, newer work demonstrates that immunotherapy can complement standard cancer treatments. With the FDA approval of two immunotherapies for clinical use, cancer immunotherapy has taken its place in the clinic as a bona fide approach to cancer treatment. The first cancer vaccine (sipuleucel-T, Provenge) was approved for advanced prostate cancer [2], and the first immune checkpoint inhibitor targeting the negative regulatory molecule cytotoxic T lymphocyte antigen-4 (CTLA-4) on T cells (ipilimumab, Yervoy) was approved for the treatment of advanced melanoma [3]. More recently, two clinical trials of antibodies that target the negative immune checkpoint molecule PD-1 on T cells and its ligand B7-H1/PD-L1 on tumor cells were unexpectedly successful, with durable response rates of 20–25 % in advanced melanoma, renal cell cancer, and nonsmall cell lung cancer (NSCLC) [4, 5]. Despite the promise of these immunologically targeted approaches to cancer treatment, a number of developmental challenges remain to be solved if we are going to realize the full potential of cancer immunotherapy. First, understanding the differences in immunobiology between the distinct histologies and biologic subtypes of cancer will be critical for identifying the optimal antigen and/or immunologic pathway to target for a particular cancer. Second, dissecting mechanisms of intrinsic and adaptive therapeutic resistance to immune-based treatments will be critical for ensuring clinical success. Finally, delineating the impact of established cancer drugs and standard cancer treatment modalities on the immune system and on tumor immunobiology itself will be critical for the most effective integration of immune-based cancer therapy into state-of-the-art multimodality cancer care. Current data suggest that combining chemotherapy in standard and novel ways with immune-based interventions will have great potential for optimizing the clinical outcomes of cancer patients.

Challenges to effective natural and therapeutic cancer immunity

An effective immune response to cancers should result in the regression of established tumors and should also be able to prevent the development of a new cancer. However, multiple factors present a barrier to the antitumor immune response. Because tumors are frequently perceived by the immune system as “self,” the mechanisms that control the development of autoreactive immune responses (and thus, autoimmune disease) also serve to preclude the development of an effective immune response to cancer (reviewed in [6]). The deletion of high-avidity autoreactive T cells during thymic education leaves in place a pool of low-avidity T cells specific for self-antigens (including tumor antigens) that are functionally suboptimal. Regulatory T cells provide a backup mechanism of regulation that helps to keep these cells and any high-avidity T cells that escape deletion from attacking both normal tissues and tumors. Furthermore, crosstalk between progressing tumors and the host immune system results in multiple superimposed mechanisms of additional regulation and immune escape that serve to keep the immune response to tumors shut down. A variety of immune cells that promote tumor growth and inhibit tumor-associated immune responses, or both, accumulate within the tumor and its locoregional draining lymph nodes. In particular, these include CD4⁺CD25⁺FOXP3⁺ regulatory T cells (Tregs), CD4⁺interleukin-17-producing T helper cells, myeloid-derived suppressor cells (MDSCs), and tumor-associated macrophages (TAMs). Additional features of the tumor microenvironment further silence the antitumor immune response, including high levels of suppressive intratumoral cytokines [transforming growth factor-β (TGF-β), tumor necrosis factor (TNF), and interleukin-10 (IL-10)], the constitutive or induced expression of immune checkpoint molecules by the tumor cells (PD-L1, B7-H4), and various other phenotypic alterations that lead to immune escape (the loss of tumor antigens and other molecules essential for antigen processing and presentation). In addition, many standard and high-dose chemotherapy regimens can be immunosuppressive, by either frankly inducing lymphopenia or contributing to lymphocyte dysfunction. Other medications used as a critical part of anticancer therapies, or given as adjunctive medications to mitigate the side effects of therapy, can also adversely affect the immune system.

Because surgery, radiotherapy, chemotherapy, and targeted therapies are widely used to treat most established cancers, the optimal integration of immune-based therapies with these standard modalities to minimize antagonistic interactions between different treatment modalities is of great importance. In addition, multiple lines of evidence suggest that chemotherapy in particular may have the potential to augment cancer immunity in several distinct ways. From a simple point of view, giving immune-based therapy in the setting of minimal residual disease—after optimal tumor debulking by surgery and/or chemotherapy—simply mitigates the negative impact of tumor burden on the potency and efficacy of the antitumor immune response. Some forms of chemotherapy-induced cell death can be highly immunogenic, and chemotherapy-induced apoptosis in particular can augment cross-priming of the tumor-specific T cell response [7]. For example, standard neoadjuvant paclitaxel therapy for early breast cancer can result in the development of new intratumoral immune cell infiltrates, with the clinical and pathologic response directly correlating with the tumor cell apoptotic response [8]. Chemotherapy may promote antitumor immunity in more subtle ways, modulating the expression of tumor antigens and immune checkpoint molecules as well as molecules more directly involved in antigen processing and presentation [reviewed in 9]. Finally, chemotherapy may impinge on mechanisms of immune tolerance and suppression to promote antigen-specific immunity [9]. A seminal study demonstrated that a single low dose of cyclophosphamide given 1–3 days before antigen exposure overcomes systemic immune tolerance to enhance both the antibody and the T cell response [9]. Conversely, concurrent or subsequent administration induced antigen-specific tolerance. It is clear that carefully choosing the chemotherapy drugs, their dose, and their timing of administration in relation to immune-based therapy is essential for engaging the potential additive and synergistic clinical activities of chemotherapy and immunotherapy, and leading to maximal clinical benefit.

Mechanisms of chemotherapy-induced immunomodulation

Chemotherapy can promote tumor immunity in two major ways: (a) through its intended therapeutic effect of killing cancer cells by immunogenic cell death, and (b) through ancillary and largely unappreciated effects on both the malignant and normal host cells present within the tumor microenvironment. A growing body of data demonstrates that some chemotherapy drugs given at their standard dose and schedule exert their antitumor effect at least in part by inducing immunogenic cell death [10]. This process can involve either the pattern recognition receptor toll-like receptor-4 (TLR-4) or the purinergic receptor P2RX7. The clinical relevance of these immunogenic cell death pathways is suggested by the association of loss of function polymorphisms of both TLR-4 and P2RX7 with a higher risk of breast cancer relapse after adjuvant anthracycline-based chemotherapy [11, 12]. Consistent with these observations, advanced stage colon cancer patients treated with oxaliplatin chemotherapy have a shorter progression-free survival (PFS) and overall survival (OS) if they harbor the TLR-4 loss of function polymorphism than if they carry the normal TLR-4 allele [13]. Notably, these same loss of function alleles seem to have no effect on clinical outcome in patients with NSCLC, suggesting that the potential for immunogenic cell death is either dependent on the tumor biology, the chemotherapy, or both [14]. In addition to immunogenic tumor cell death, there are a number of ancillary mechanisms by which chemotherapy may promote the antitumor immune response [9]. These include enhanced intrinsic tumor cell immunogenicity, the modulation of a variety of immunoregulatory cells [including Tregs and other CD4⁺ T cells, MDSCs, and dendritic cells (DCs)], and the induction of homeostatic proliferation.

Enhanced tumor cell immunogenicity

Various chemotherapy agents can modulate intrinsic tumor cell immunogenicity in multiple ways (summarized in Table 1). Chemotherapeutics can upregulate the expression of tumor antigens (CEA and cancer–testis antigens, for example) as well as MHC class I molecules, thereby increasing capacity for tumor antigen presentation. Some drugs induce the expression of co-stimulatory molecules such as B7-1 on the surface of tumor cells, rendering them better able to present tumor antigens to achieve effective immune activation. Other drugs decrease the expression of immune checkpoint molecules such as B7-H1/PD-L1 and B7-H4 on the tumor cell surface, thereby preventing infiltrating T cells from being shut down. Finally, other chemotherapy drugs can sensitize tumor cells to CD8⁺ T cell-mediated lysis through fas-, perforin-, and granzyme-B-dependent mechanisms.

Table 1.

Modulation of tumor cell immunogenicity by chemotherapy drugs

Mechanism	Chemotherapy drugs	Tumor cell type	Reference
Sensitizes to TRAIL-dependent CD8+ T cell-mediated apoptosis	Cyclophosphamide	Mesothelioma cells	[15]
Sensitizes tumor cells to CD8+ T cell-mediated apoptosis through fas-, perforin-, or granzyme-mediated pathways	5-Fluorouracil Dacarbazine	Melanoma	[16]
Upregulation of mannose-6-phosphate receptors and increased permeability to granzyme-B released by CTLs	Paclitaxel Cisplatin Doxorubicin	Lymphoma Colon cancer cells	[17]
Induce autophagy, resulting in tumor-infiltrating DCs and T cells by released ATP	Mitoxantrone Oxaliplatin	Colon cancer cells	[18]
Induce immunogenic cell death through cell surface expression of calreticulin	Gemcitabine Cyclophosphamide Oxaliplatin Doxorubicin Paclitaxel	Colon cancer, melanoma, and fibrosarcoma cells	[19]
Increases expression of MHC class I and antigen-processing machinery components	Paclitaxel	Colon cancer cells	[20]
Expression of carcinoembryonic antigen (CEA)	5-Fluorouracil	Colon and breast cancer cells	[21]
Downregulation of B7-H1/PD-L1 from cell surface	Doxorubicin Ara-C (cytosine arabinoside)	Breast cancer and leukemia cells	[22, 23]
Increased expression of cell surface B7-1 and B7-2	Ara-C (cytosine arabinoside)	Leukemia cells Mastocytoma and plasma-cytoma cells	[23–25]
Upregulate expression of cancer–testis antigens	5′-aza-2′deoxycytidine	Renal cell carcinoma, ovarian cancer, melanoma, glioma	[26, 27]

TRAIL tumor necrosis factor-related apoptosis-inducing ligand, CTLs cytotoxic T lymphocytes, DCs dendritic cells, ATP adenosine triphosphate, MHC major histocompatibility complex, CEA carcinoembryonic antigen, PD-L1 programmed death ligand-1, Ara-C cytosine arabinoside

Regulatory T cells and other CD4⁺ T cells

CD4⁺ T cells can be classified into several distinct subsets, including Tregs, CD4⁺ T helper type 1 cells, CD4⁺ T helper type 2 cells, and CD4⁺ T helper type 17 cells [28]. The fact that CD4⁺ T cells are characterized by considerable phenotypic and functional plasticity complicates their classification, but it is clear that some chemotherapy drugs can promote or suppress functions attributed to each of these subsets.

Tregs represent 5–10 % of the peripheral CD4⁺ T lymphocyte pool and play a key role in maintaining peripheral tolerance and preventing autoimmunity by inhibiting CD8⁺ T cell responses in an interleukin-2-dependent manner [29]. These cells expand in the peripheral blood and within the tumors of patients with breast, pancreas, ovary, stomach, lung, and liver cancers [30]. There, they potently suppress the antitumor immune response. Many groups have shown in both preclinical models and in cancer patients that several chemotherapy drugs can abrogate the suppressive influence of Tregs and allow effective antitumor immunity to emerge (Table 2). Both cyclophosphamide and cisplatin can decrease Tregs and enhance antigen-specific CD8⁺ T cell activity in murine models [31–37]. Also in murine models, the metronomic administration of cyclophosphamide, paclitaxel, or temozolomide can decrease Treg activity [38–40]. Clinically, cyclophosphamide at doses of 200–300 mg/m² given 1 day prior to vaccination and 600 mg/m² given 7 days prior to vaccination has been shown to decrease Tregs [30]. In addition, metronomic cyclophosphamide alone can deplete Tregs and restore T and NK cell effector function in advanced cancer patients [41]. Together, these findings illustrate the importance of dose, schedule, and therapeutic intent in determining the best administration schema for cyclophosphamide. In addition, fludarabine- or paclitaxel-based chemotherapy can decrease Tregs in cancer patients [42–44], an effect sometime associated with improved clinical response [45]. In addition, several chemotherapy drugs can shift the CD4⁺ T helper cell phenotype from type 2 to type 1, thereby reversing immunologic skew and promoting tumor immunity [9, 36]. Finally, cyclophosphamide can also promote the CD4⁺ T helper type 17 cell phenotype and function [46, 47].

Table 2.

Modulation of immunoregulatory cells by chemotherapy drugs

Mechanism	Chemotherapy drugs
Decreased Treg number or function	Low dose
	Cyclophosphamide
	Standard dose
	Cyclophosphamide
	Paclitaxel
	Fludarabine
	Cisplatin
	Metronomic
	Cyclophosphamide
	Paclitaxel
	Temozolamide
Promote T helper type 1 immunity	Cyclophosphamide
	Paclitaxel
	Bleomycin
	Melphalan
Promote T helper type 17 immunity	Cyclophosphamide
Increase MDSCs	Cyclophosphamide
	AC chemotherapy
Decrease MDSCs	Gemcitabine
	5-Fluorouracil
	Cisplatin
	Docetaxel
Enhance DC function	Cyclophosphamide
	Vincristine
	Vinblastine
	Paclitaxel
	Docetaxel
	Methotrexate
	Mitomycin-C
	Doxorubicin
	5′-aza-2′-deoxycytidine

Treg = CD4⁺CD25⁺FoxP3⁺ regulatory T cells

MDSC myeloid-derived suppressor cells, AC doxorubicin/cyclophosphamide chemotherapy, DCs dendritic cells

Myeloid-derived suppressor cells (MDSCs)

MDSCs are a heterogeneous population of myeloid cells that progressively accumulate in cancer patients and potently inhibit T cell-mediated immune responses through the production of nitric oxide and arginase [48]. In murine models, cyclophosphamide administration increases peripheral MDSC numbers in tumor-free mice [49]. In patients, MDSC numbers increase in breast cancer patients treated with standard doxorubicin and cyclophosphamide (AC) chemotherapy [50]. Standard doses of gemcitabine, 5-fluorouracil, and cisplatin can decrease peripheral MDSC in mice, thereby enhancing the potential for expanding CD8⁺ T cells in vitro and in vivo, both naturally and after vaccination [37, 51, 52]. Docetaxel decreases MDSC and increases CD8⁺ T cell activity in a murine model of breast cancer [53]. Specifically, it selectively decreases regulatory MDSC while sparing other MDSC and polarizing them toward a functional antitumor phenotype. These findings together suggest distinct chemotherapy drugs impact MDSCs in idiosyncratic ways.

Dendritic cells (DCs)

Multiple chemotherapy drugs given at low doses can enhance the function of DCs; these include cyclophosphamide, vincristine, vinblastine, paclitaxel, docetaxel, methotrexate, mitomycin-C, doxorubicin, and 5-aza-2′-deoxycytidine [54, 55]. Low-dose cyclophosphamide differentially decreases the bone marrow-derived CD8⁺ tissue-resident DC relative to migratory or plasmacytoid DC, thereby increasing effective antigen presentation and cytokine secretion [56]. It also promotes DC maturation [57] and induces the production of type 1 interferons [58], ultimately resulting in the evolution of a durable CD44^hi memory T cell response. Nonmyeloablative doses of cyclophosphamide induce compensatory myelopoiesis, which promotes the trafficking of DC to tumors [59]. There, they secrete high levels of IL-12 relative to IL-10, favoring the function of effector T cells relative to Tregs. Lymphodepleting doses of cyclophosphamide cause the proliferation of DC within the bone marrow, followed by their expansion in the periphery [60]. These DCs are capable of effectively presenting antigens to T cells [61].

The taxanes also impact DC phenotype and function. Low-dose paclitaxel facilitates DC maturation and proinflammatory cytokine secretion (particularly IL-12) in a TLR-dependent fashion [62–64]. This results in augmented CD8⁺ T cell priming and cytotoxic activity [65, 66]. In vivo, subclinical doses of paclitaxel given 1 day prior to a human epidermal growth factor receptor-2 (HER-2)-targeted granulocyte and macrophage colony-stimulating factor (GM-CSF)-secreting tumor vaccine shifts the CD4⁺ T helper response from type 2 to type 1 in mice immunologically tolerant to HER-2, delaying tumor outgrowth [36]. Combining paclitaxel and cyclophosphamide at low doses synergistically delays the outgrowth of HER-2⁺ tumors in this model [65]. Both paclitaxel and docetaxel at low doses can augment the activity of other vaccine platforms in other preclinical models, enhancing tumor-specific immune responses and tumor-free survival [67–70].

Homeostatic proliferation

A growing body of evidence suggests that the simple presence of lymphopenia may be associated with worse clinical outcomes [71]. Conversely, lymphopenia induced with therapeutic intent can create space for reshaping the repertoire of immune cells available for fighting cancer, with the potential for improving clinical outcomes [72, 73]. Strategically inducing significant lymphopenia creates both space and a cytokine milieu that includes high levels of IL-7, IL-15, and IL-21, promoting the proliferation of T cells that recognize self [74]. This proliferating T cell repertoire can be further skewed toward a desired antigenic specificity by vaccination or by the adoptive transfer of antigen-specific lymphocytes during homeostatic proliferation [75]. In a preclinical lung cancer model, the chemotherapy doublet cisplatinum/vinorelbine induces a lymphopenia that promotes the vaccine-induced expansion of effector T cells relative to Tregs through the influence of IL-7 [76]. Vaccine-induced tumor immunity can also be augmented when tumor-bearing mice are vaccinated during early engraftment after either syngeneic or allogeneic T cell-depleted bone marrow transplantation; it is further enhanced by the infusion of donor leukocytes from vaccinated donor mice [77–79]. These strategies have been tested in clinical trials of adoptive T cell transfer alone or combined with vaccination during immune reconstitution with some success [80–82]. In contrast, after induction timed sequential therapy for newly diagnosed acute myelogenous leukemia (AML), recovering T lymphocytes were predominantly peripherally derived, phenotypically activated, and functionally suppressive Tregs, where the expansion was antigen-driven [83]. It is clear that optimizing treatment regimens that induce lymphopenia, and characterizing the kinetics and longevity of T cell recovery as well as the phenotype and functional characteristics of lymphocyte recovery, will be essential to harness the power of homeostatic proliferation for targeted tumor therapy.

Clinical trials of chemoimmunotherapy

Clinical trials incorporating cancer vaccines, immune checkpoint blockade, or adoptive cellular therapy have typically tested the immunotherapies integrated with standard-dose chemotherapy, sequenced with low-dose chemotherapy specifically designed for immune modulation, or given in the setting of high-dose chemotherapy designed to reboot the immune system through cytokine release and homeostatic proliferation. These clinical trials are summarized in Tables 3 and 4.

Table 3.

Clinical trials of chemoimmunotherapy with vaccines

Patient population	n	Vaccine	Chemotherapy drug	Immunologic outcome
Concurrent standard-dose chemotherapy
Stage 2/3 pancreas cancer (ref [85])	60	GM-CSF-secreting, mesothelin+ pancreas tumor cells	Vaccine once, then 5-fluorouracil-based chemoradiation, then 3 more vaccines	Induction of mesothelin-specific T cells, inhibited by chemoradiation, then restored by 3 vaccines
Stage 4 prostate cancer (ref [86, 87])	405	GM-CSF-secreting prostate tumor cells	Docetaxel 75 mg/m2 + vaccine versus + dexamethasone	Not reported, study terminated
Stage 4 colon cancer (ref [88])	118	ALVAC-CEA-B7.1	Standard 5-fluorouracil, leucovorin, irinotecan	No inhibition of CEA-specific immunity
Stage 4 colon cancer (ref [89])	17	CEA peptide	Standard 5-fluorouracil, leucovorin, irinotecan	50 % patients with new CEA-specific T cells
Stage 4 prostate cancer (ref [90])	28	PSA vaccinia/fowl pox	Standard docetaxel and dexametha-sone	No inhibition of PSA-specific immunity
Stage 4 melanoma (ref [91])	25	GV1001 telomerase peptide	Temozolomide 200 mg/m2 daily ×5	78 % patients with new, durable GV1001-specific immunity
Stage 3 NSCLC (ref [92])	23	GV1001 telomerase peptide	Weekly docetaxel 20 mg/m2 + XRT,	80 % patients with new GV1001-specific immunity, 65 % durable
Stage 4 melanoma (ref [94, 95])	10	Melan-A/MART-1/gp-100 peptide	No dacarbazine versus dacarbazine 800 mg/m2 (n = 5 in each group)	Enhanced peptide-specific immunity with dacarbazine
Stage 4 cancers (ref [96])	24	Type 1 DCs + autologous tumor lysate + GM-CSF + peg-IFN	CY 600 mg/m2 D-7	36 % with new immunity
Low-dose immunomodulatory chemotherapy
Stage 4 breast cancer (ref [106])	1,028	KLH (n = 505) KLH-STn (n = 523)	CY 300 mg/m2 each arm D-3	Not reported, no clinical difference
Stage 4 pancreas cancer (ref [107])	50	GM–CSF-secreting, mesothelin+ pancreas tumor cells	No CY versus CY 300 mg/m2 D-1	Enhanced mesothelin-specific T cells with CY 300 mg/m2
Stage 2–4 ovarian cancer (ref [109])	11	Peptide-loaded DCs: HER-2, hTERT, PADRE	No CY versus CY 300 mg/m2 D-2	Modest T cell responses to HER-2 and hTERT
Stage 4 renal cell carcinoma (ref [110])	68	Multipeptide vaccine IMA901	No CY versus CY 300 mg/m2 D-3	Enhanced peptide-specific responses on each arm, no difference in magnitude
Stage 4 breast cancer (ref [111])	28	GM–CSF-secreting, HER-2+ breast tumor cells	CY 0, 200, 250, 350 mg/m2 D-1 DOX 0, 15, 25, 35 mg/m2 D-7 Factorial response surface design of various dose combinations	Increased HER-2-specific immunity with doses of CY 200 mg/m2 + DOX 35 mg/m2

ALVAC = canary pox virus; XRT = radiation therapy; MART-1 = melanoma antigen recognized by T cell; DCs = dendritic cells; STn = clustered carbohydrate antigen

GM–CSF granulocyte–macrophage colony-stimulating factor, CEA carcinoembryonic antigen, PSA prostate-specific antigen, NSCLC nonsmall cell lung cancer, IFN interferon, KLH keyhole limpet hemocyanin, D day, CY cyclophosphamide, DOX doxorubicin, hTERT human telomerase reverse transcriptase, PADRE pan DR-binding peptide

Table 4.

Clinical trials of chemotherapy combined with immune checkpoint blockade

Patient population	n	Immune checkpoint modulator	Chemotherapy drug	Clinical outcome
Stage 4 metastatic melanoma (ref [93])	502	Ipilimumab (n = 251) Placebo (n = 251)	DTIC 850 mg/m2 each arm	Improved OS, and 1, 2, and 3 year survival with Ipilimumab + DTIC versus DTIC alone
Stage 4 NSCLC (ref [97])	204	Ipilimumab, concurrent versus phased	Paclitaxel 175 mg/m2 Carboplatin AUC 6	Improved immune-related and standard PFS with phased treatment with chemotherapy first, followed by chemotherapy with ipilimumab
Extensive-stage SCLC (ref [98])	130	Ipilimumab, concurrent versus phased	Paclitaxel 175 mg/m2 Carboplatin AUC 6	Improved immune-related but not standard PFS with phased treatment with chemotherapy first, followed by chemotherapy with ipilimumab
Stage 4 breast cancer (ref [99])	30	IMP321	Paclitaxel 80 mg/m2 D1, 8, 15 weekly	Clinical benefit of 90 % Objective response rate of 50 % versus 25 % for historical control group Increased APC number and activity, increased NK percentage, increased CD8+ memory T cells

DTIC dacarbazine, IMP321 recombinant soluble LAG-3Ig fusion protein

OS overall survival, NSCLC nonsmall cell lung cancer, AUC area under the curve, PFS progression-free survival, SCLC small cell lung cancer, APC antigen-presenting cell, NK natural killer cell

Standard-dose chemotherapy and immunotherapy

Many standard-dose chemotherapy regimens suppress the antitumor immune response by directly decreasing lymphocyte numbers or inhibiting lymphocyte functions. Glucocorticoids are a critical component of many chemotherapy regimens, compounding the immune suppressive effects of cytotoxic therapies through their direct lymphocytic effects. However, certain chemotherapy agents can generate tumor-specific immune responses as part of their mechanism of action, primarily by inducing immunogenic cell death. This duality underlies the observation that standard-dose chemotherapy can repress vaccine-induced immune responses, augment vaccine-induced immune responses, or have no impact on vaccine activity. In part, the impact of chemotherapy on vaccine-induced tumor immunity depends on its timing of administration in relation to vaccination.

Concurrent standard-dose chemotherapy may inhibit immunotherapy

One study examined immune responses to vaccination with the canary pox vaccine ALVAC-CEA-B7.1 in patients with advanced carcinoembryonic antigen (CEA)-expressing malignancies [84]. A retrospective analysis showed that a greater number of prior chemotherapy regimens and more recent standard-dose cytotoxic therapy were both associated with lower numbers of vaccine-induced CEA-specific T cells. A phase II study integrated a mesothelin-expressing, GM-CSF-secreting pancreas tumor vaccine with standard therapy for stage II and III pancreas cancers [85]. Participants developed mesothelin-specific T cell responses after one vaccination given prior to pancreaticoduodenectomy and then went on to receive 6 months of 5-fluorouracil-based chemoradiation, with ensuing loss of the mesothelin-specific T cell response. Three additional vaccinations given after adjuvant chemoradiation restored the mesothelin-specific T cell response, suggesting that surgery and chemoradiation had suppressed vaccine-induced immunity. A phase III study tested a GM-CSF-secreting prostate tumor vaccine with standard-dose docetaxel chemotherapy and randomized patients to vaccination with ten cycles of docetaxel every 3 weeks without prednisone or ten cycles of docetaxel every 3 weeks with prednisone 10 mg daily [86, 87]. Due to an imbalance of deaths (all due to disease progression and death from prostate cancer) on the experimental arm, the study was terminated prior to completion. With subsequent follow-up, the imbalance of deaths lessened from 20 to 9. Two major factors likely contributed to the failure of this trial. First, the standard chemotherapy regimen for advanced prostate cancer includes both docetaxel and prednisone, but the prednisone was not given to vaccinated patients due to concern for immune suppression. Second, the vaccine itself had never been tested in the setting of chemotherapy before. This dose and schedule of docetaxel may have impaired vaccine activity.

Concurrent standard-dose chemotherapy may facilitate immunotherapy

In one study, a CEA-specific peptide vaccine was given first concurrently with 3 cycles of high dose 5-fluorouracil/leucovorin plus standard dose irinotecan and then alone in 17 patients with metastatic colorectal cancer at first relapse. There were 6 objective responses, and 5 patients with stable disease and 6 patients with progressive disease [88]. Immunologically, about 50 % of participants developed CEA-specific T cell responses, whereas during active chemotherapy, recall CD8⁺ T cell immunity to Epstein–Barr virus (EBV) and cytomegalovirus (CMV) declined by about 14 %. The impact of chemotherapy (5-fluorouracil, leucovorin, and irinotecan) on immunity induced by the ALVAC-CEA-B7.1 vaccine was further explored prospectively in 118 patients with metastatic colorectal cancer [89]. Three groups were treated with (a) vaccine alone for 3 cycles followed by vaccine and concurrent chemotherapy; (b) vaccine with tetanus toxoid adjuvant given for 3 cycles followed by vaccine and concurrent chemotherapy; or (c) 4 cycles of chemotherapy followed by 4 cycles of vaccination in patients without disease progression on chemotherapy. Although no significant differences in clinical or immunologic responses between the three groups emerged, CEA-specific T cells increased between 30 and 50 % across the groups, suggesting that chemotherapy did not impair vaccine-induced immunity. Another study tested a prime/boost recombinant vaccinia virus (rVV)/fowl pox (rF) vaccination regimen for patients with metastatic hormone-refractory prostate cancer [90]. Patients were vaccinated with rVV expressing prostate-specific antigen (PSA) admixed with rVV expressing the co-stimulatory molecule B7.1, followed by vaccination with rF-PSA alone or with concurrent weekly docetaxel (30 mg/m²). PSA-specific T cells increased by greater than threefold over 3 months regardless of concurrent docetaxel therapy, and median progression-free survival (PFS) was 6.1 months compared to 3.7 months for a historical patient cohort treated with weekly docetaxel. Another trial in 25 patients with metastatic melanoma showed that the telomerase-specific peptide vaccine GV1001 can be combined with temozolomide at 200 mg/m² orally for 5 days every 4 weeks, with the GV1001 vaccinations being given as 8 injections over 11 weeks [91]. The majority (78 %) of patients developed a GV1001-specific immune response characterized by high IFN-γ/IL-10 ratios, polyfunctional cytokine profiles, and the capacity to recognize naturally processed antigens. Furthermore, long-term immunologic memory was associated with clinical benefit. Another trial tested concurrent chemoradiation (2 Gray for 30 days and weekly docetaxel 20 mg/m²) and GV1001 vaccination in 23 patients with inoperable stage III NSCLC [92]. Here, GV1001-specific immunity was observed in 16 of 20 evaluable patients (80 %), with durable responses in 13 subjects. An immune response was associated with a median PFS of 371 days, compared with 182 days for immune nonresponders. Finally, a seminal paper demonstrated that concurrent therapy with dacarbazine and ipilimumab was associated with improved overall survival (OS) in patients with untreated metastatic melanoma [93].

Sequencing standard-dose chemotherapy before or after immune-based therapy may augment immunotherapy

Several studies have examined phased treatment with standard-dose chemotherapy and immunotherapies, where most studies explored giving chemotherapy first. A study of 36 patients with stage 2–4 melanoma and no evidence of disease were treated either with 800 mg/m² dacarbazine (DTIC) 1 day prior to vaccination with melan-A/MART-1/gp-100 peptides, or with the peptide vaccine alone [94]. The administration of DTIC prior to vaccination induced the expression of genes involved in the immune response, increased numbers of peptide-specific CD8⁺ T cells, and promoted the generation of a long-lived memory T cell response. Further analyses also revealed a broader T cell receptor diversity, higher T cell avidity, greater tumor reactivity, and a trend toward longer survival in DTIC-pretreated patients [95]. Finally, a pilot clinical trial of 24 patients with metastatic cancers tested type 1 DCs loaded with autologous tumor lysates given as 2 cycles of 4 daily vaccinations sequenced with cyclophosphamide at 600 mg/m² 7 days prior to vaccination, followed by GM-CSF days 1–4 after vaccination and pegylated IFN-α days 1 and 8 after vaccination [96]. This study showed that cyclophosphamide pretreatment reduced Tregs to levels observed in healthy individuals. T cell responses specific for DC+ tumor lysate were observed in 4 of 11 tested patients, and 5 patients demonstrated disease stabilization.

Immune checkpoint blockade has also been tested with standard-dose chemotherapy (Table 4). A study in NSCLC and extensive-stage SCLC examined the use of concurrent standard chemotherapy with 4 cycles of carboplatin and paclitaxel plus the immune checkpoint inhibitor ipilimumab followed by 2 cycles of carboplatin and paclitaxel as compared to phased treatment with 2 cycles of standard carboplatin and paclitaxel followed by 4 cycles of ipilimumab plus carboplatin and paclitaxel [97, 98]. In both patient groups, improved immune-related PFS was observed for phased ipilimumab, where standard chemotherapy was given first for two cycles before ipilimumab was added. Finally, a study in 30 patients with metastatic breast cancer tested the immune checkpoint modulator IMP321 (recombinant soluble LAG3Ig fusion) every 2 weeks preceded by standard dose weekly paclitaxel [99]. This study showed evidence of both clinical benefit, with objective response rates of 50 % compared to historical controls, and immune activation, with increases in natural killer cells as well as durable effector-memory CD8⁺ T cell responses.

Some evidence suggests that patients treated with immune-based therapy and who subsequently require salvage chemotherapy may have an enhanced clinical response. One study of DCs transduced with a p53-expressing adenoviral vector in 29 patients with extensive-stage SCLC revealed that 60 % of patients developed p53-specific immunity, but all patients except one progressed on study [100]. Notably, about 60 % of patients with progressive disease—those with p53-specific immunity–demonstrated an objective response to salvage chemotherapy. Similarly, glioblastoma patients vaccinated with a dendritic cell vaccine who progressed on study also had a better than expected response to postvaccination salvage chemotherapy [101].

High-dose chemotherapy and immunotherapy

Several groups have also explored high-dose chemotherapy given with or without radiation as a platform for adoptive cellular therapy or vaccination. T cells specific for minor histocompatibility antigens were administered after allogeneic stem cell transplantation in one small 7 patient study [102]. The transferred T cells persisted for up to 21 days, and 5 of the 7 patients treated demonstrated complete but temporary remissions after adoptive T cell transfer. A larger clinical trial tested two types of conditioning followed by adoptive cellular therapy with autologous tumor-infiltrating lymphocytes (TIL) and IL-2 in 50 patients with metastatic melanoma [103]. High-dose cyclophosphamide and fludarabine alone resulted in a response rate of 49 %, and adding 2 or 12 Gray of total body irradiation resulted in response rates of 52 and 72 %, respectively. High serum levels of IL-7 and IL-15 were associated with lymphodepletion, and objective tumor responses correlated with longer telomere length of the transferred lymphocytes. High-dose chemotherapy followed by autologous stem cell transplantation as a platform for the administration of a GM-CSF-secreting, cell-based autologous tumor vaccine has been tested in at least 2 clinical trials that enrolled patients with acute myelogenous leukemia (AML), with an acceptable safety profile and evidence of vaccine-induced immune responses [82, 104].

Low-dose chemotherapy and immunotherapy

Immune modulating doses of chemotherapy (typically much lower than the standard doses used for cancer treatment) can be used to strategically enhance the activity of tumor vaccines in clinical trials rather than to kill cancer cells directly (Table 3). Several phase II clinical trials are historically important as they investigated the use of low-dose cyclophosphamide to relieve the inhibitory influence of suppressor T cells as they were defined in the 1970s and 1980s. In the aggregate, these early studies demonstrated that 300 mg/m² cyclophosphamide given 3 days prior to vaccination with a clustered carbohydrate antigen-keyhole limpet hemocyanin (KLH) vaccine resulted in higher antibody titers and longer survival than vaccination without cyclophosphamide pretreatment [105]. To definitely test these findings, a phase III clinical trial testing cyclophosphamide-modulated vaccination for metastatic breast cancer randomized 505 patients to treatment with 300 mg/m² cyclophosphamide followed by vaccination with KLH alone and 523 patients to treatment with 300 mg/m² cyclophosphamide followed by vaccination with clustered carbohydrate-KLH [106]. Although PFS and OS were no different for the two groups, an unplanned subset analysis suggested a possible trend toward improved PFS in patients in concurrent endocrine therapy.

The use of low-dose cyclophosphamide for immune modulation has subsequently been tested in several clinical trials. One study tested a GM-CSF-secreting, mesothelin-expressing cell-based tumor vaccine alone or given 1 day after cyclophosphamide at 300 mg/m² in 50 patients with metastatic pancreas cancer [107]. Although no definitive impact of chemotherapy was observed, there was a trend toward both increased mesothelin-specific CD8⁺ T cell responses and clinical benefit with cyclophosphamide-modulated vaccination as compared to vaccination alone.

A similar vaccine platform was given with 300 mg/m² cyclophosphamide to patients with advanced NSCLC, with a transient decrease in peripheral Tregs with time after cyclophosphamide observed [108]. A third study reported the activity of a peptide vaccine consisting of monocyte-derived DCs loaded with peptide epitopes specific for HER-2, hTERT, and PADRE alone or with 300 mg/m² cyclophosphamide given 2 days prior to vaccination in 11 patients with advanced ovarian cancer in remission [109]. The patients treated with cyclophosphamide-modulated vaccination had a trend toward improvement in survival compared to those treated with vaccination alone, and the 3-year OS for this small group of patients was 90 %. Adding cyclophosphamide to vaccination was associated with a transient decrease in neutrophils, but no change in peripheral lymphocytes or Tregs was observed. A recent randomized phase II clinical trial conducted in 68 patients with advanced renal cell carcinoma tested 17 vaccinations with a multipeptide vaccine (IMA901), where the first vaccination was given 3 days after single dose of 300 mg/m² cyclophosphamide in 33 patients, and the 17 vaccinations were given with no pretreatment in a separate group of 35 patients [110]. Although there was no difference in PFS between the groups, there was trend toward longer OS in patients who received cyclophosphamide-modulated vaccination compared to those who received vaccination alone (median OS of 23.5 vs. 14.8 months). Interestingly, the immune response rates between the two groups were comparable, suggesting that cyclophosphamide pretreatment did not influence the induction of T cell responses. Furthermore, among immune responders, those pretreated with a single dose of cyclophosphamide had longer OS than those who were not pretreated with cyclophosphamide (HR 0.38, p = 0.01); there was no difference in survival between the arms in immune nonresponders (HR = 0.92, p = 0.870). In addition, those who developed multipeptide immune responses survived longer than those who did not, suggesting an association of clinical benefit with a diverse immune response. The single dose of cyclophosphamide resulted in a roughly 20 % decline in Tregs at 3 days as compared to baseline. The percentage of proliferating cells among all regulatory T cells was also decreased. There was no impact on absolute lymphocyte counts on either arm, and these effects were not observed in patients who received vaccination without cyclophosphamide. A distinct study tested a HER-2-positive, GM-CSF-secreting, cell-based breast tumor vaccine alone, or given in a specifically timed sequence with low, immune modulating doses of the two chemotherapy drugs cyclophosphamide and doxorubicin [111]. An innovative factorial response surface design was used to detect possible interactions between cyclophosphamide and doxorubicin, and the cyclophosphamide was given at doses of 0, 200, 250, or 350 mg/m² 1 day prior to vaccination, whereas the doxorubicin was given at doses of 0, 15, 25, or 35 mg/m² 7 days after vaccination. Vaccination alone induced new HER-2-specific delayed type hypersensitivity (DTH), with low levels of HER-2-specific antibodies also detected by ELISA. The addition of cyclophosphamide at 200 mg/m² maintained the DTH response and enhanced the HER-2-specific antibody response, whereas doses of cyclophosphamide at 250 or 350 mg/m² abrogated vaccine-induced immunity. Preliminary data suggest a decrement in peripheral Tregs that is dependent of the dose of cyclophosphamide. The chemotherapy dose combination that maximized vaccine-induced HER-2-specific antibody responses was 200 mg/m² cyclophosphamide and 35 mg/m² doxorubicin. Taken together, these data suggest there may be a narrow therapeutic window for the use of low-dose cyclophosphamide for immune modulation, particularly when used in sequence with low-dose doxorubicin.

Conclusions

A new era of effectively harnessing the immune system to treat and prevent cancer has begun. Two distinct immune-based therapies are now approved for cancer treatment: the first cancer vaccine (sipuleucel-T for advanced prostate cancer), and the first immune checkpoint inhibitor (ipilimumab for advanced melanoma). These early successes have led to heightened interest and activity in developing new strategies for tipping the balance of the host-tumor interaction toward definitive tumor rejection. It is clear that strategically integrating immune-based therapies with standard cancer treatment modalities, in particular chemotherapy drugs, has the potential to reengineer the overall host milieu and the local tumor microenvironment to disrupt pathways of immune tolerance and suppression. In designing combination immunotherapy regimens, clinical investigators should consider how chemotherapy impacts the immune system in order to guide the dose and schedule for integrating chemotherapy and immunotherapy. In particular, systematically defining the optimal drug dose and timing in relation to immune-based therapy in early-phase clinical studies is imperative for the design of phase II and III clinical trials with a higher likelihood of clinical success. Such thoughtful clinical development plans will accelerate the clinical development of chemoimmunotherapy regimens that result in a vigorous and sustained antitumor immune response that eradicates cancer and prevents its recurrence.