**3. Effect of immune checkpoint blockade on immune suppressive T cells**

#### **3.1 Effect of PD-1 blockade on immune suppressive T cells**

Regulatory T cells (Tregs), an immunosuppressive subset of T cells, are known to be closely involved in the regulation of the immune responses to cancer [65–67]. The tumor-infiltrating subsets of Tregs are characterized by their high surface expression of PD-1 [68–70] and the PD-1/PD-L1 axis is known to modulate Tregs function via cell-intrinsic pathways. For example, the blocking of PD-1 in animal models reduced the immunosuppressive function of Tregs and declined their expression in the TME [71, 72]. Moreover, studies on murine models have shown that PD-1/PD-L1 pathway mediates the conversion of CD4<sup>+</sup> Th1 effector T cells into induced Foxp3<sup>+</sup> Tregs (iTregs) [73, 74]. Conversely, certain preclinical studies show that anti-PD-1 therapy is associated with an increase rather than a decrease in Tregs infiltration in the TME [24, 75]. The proliferation of Tregs under anti-PD-1 therapy could be explained by a treatment-induced reversal of the exhausted state of PD-1Hi in the TME [76, 77]. The expansion of Tregs upon blockade of PD-1/PD-L1 axis has been observed at the tumor level as well as in the peripheral blood. One particular study showed that patients with gastric adenocarcinoma responding to anti-PD-1 displayed an on-treatment decrease in intra-tumoral Tregs, whereas non-responders had post-treatment tumor biopsies exhibiting an infiltration of highly proliferative effector Tregs (Foxp-3hi/CD45− CD4+ ) [78]. This suggests that on-treatment changes in intra-tumoral T reg infiltration could be a relevant dynamic parameter to account for in predicting anti-PD-1/PD-L1 treatment response. The predictive insight provided by the on-treatment dynamics of circulating Tregs under PD-1/ PD-L1 blockade has also been investigated. In a study conducted on melanoma patients exposed to nivolumab, an expansion of circulating Tregs under therapy was shown to positively correlate with treatment benefit [79]. These observations suggest that Tregs dynamics under anti-Pd-1 therapy may contribute predictive insight into treatment benefit when monitored both in the TME and in the periphery.

Interestingly, another immunosuppressive CD4+ T cell subset found to be regulated by anti-PD-1 therapy has been recently identified. These cells known as 4PD-1Hi express high levels of PD-1 while lacking Foxp-3 expression [80]. Also, these 4PD-1Hi cells were shown to accumulate in the tumor and to inhibit T cells effector

*Advances in Precision Medicine Oncology*

T cells and successful anti-tumor response [33] despite similar CD8+

In addition, a tumor microenvironment that is rich in CD4+

/CD4+

and CD8+

is the enhanced expansion of memory CD8+

*Effect of immune checkpoint inhibition on effector T cells.*

CD4<sup>+</sup>

**Figure 1.**

this phenotype of CD4+

presence of CD4+

reported an increase of ICOS+

tion of both peripheral CD4+

of immunosuppressive elements in the TME, highlight that adequate CD8+

profiles observed in lesions that responded to anti-CTLA-4 therapy and those that did not respond to this therapy [33]. These findings, possibly explained by the action

long-term tumor control post-therapy. Moreover, this expansion of memory CD8+

treatment are found to exhibit a Th1-like effector phenotype with a noticeable expression of ICOS, a known marker of follicular T helper cells [30]. Interestingly,

negative co-inhibitory molecules such as CTLA-4 [38]. Moreover, ICOS-deficient mice showed an impaired anti-tumor T cell response to anti-CTLA-4 therapy [39].

infiltration was shown to be critical to develop response to anti-CTLA-4 therapy in castration resistant prostate cancer [40]. Furthermore, a higher expression of Th1 associated genes was observed in melanoma tumors of patients responding to ipilimumab compared to non-responders, supporting the functional relevance of a Th1- response in CTLA-4 inhibitor treatment benefit [41]. Interestingly, while the

blood profile rich in Th17 cells was reported to be rather predictive of autoimmune toxicity under anti-CTLA-4 therapy [42]. Moreover, several clinical studies have

patients treated with anti-CTLA-4 [43–49]. Additionally, an increased prolifera-

the first dose of anti-CTLA-4 treatment [50–52]. Such a response may partly be due to the bulk expansion in the periphery of specific T cells against known tumor antigens in anti-CTLA-4 treated patients [53–55]. Of note, an increase in absolute

cells is considered as an indicator of treatment benefit in a few clinical studies [36, 37].

T cells expanding in the tumors of mouse models under anti-CTLA-4

T cell was observed in mice after genetic knock-down of

Th1 cells is predictive of response to anti-CTLA-4, a peripheral

T cells in the tumor and peripheral blood of

T cells is observed as early as 3 weeks after

cells function is necessary but not sufficient for complete suppression of tumor growth under immunotherapy [34]. Another interesting effect noted upon CTLA-4 blockade

T cells activation

T cell [35] which will help to promote

Th1 effector T cells

effector T

T

**32**

function. Therefore, 4PD1Hi cells are considered as a marker of tumor progression. Anti-PD-1 treatment was shown to reduce the proliferation of 4PD-1Hi cells, whereas anti-CTLA-4 treatment showed an exactly opposite effect on these cells. In line with this interesting observation, the downregulation of 4PD-1Hi cells under anti-PD-1 treatment was further documented as a biomarker of treatment response under anti-PD-1 pembrolizumab antibody in a melanoma patient cohort. In addition, some preclinical studies support the fact that anti-PD-1 therapy can induce the expansion of specific CD8+ T cells immunosuppressive subsets [30]. Indeed, a recent study showed that PD-1 blockade in sub-optimally primed T cell conditions supported the proliferation of dysfunctional immunosuppressive CD8+ T cells expressing PD-1 and high levels of CD38 and this effect was associated with treatment failure and tumor resistance in cancer patients [81]. The dynamics of these T cell subsets under treatment may therefore provide valuable predictive information, pending the validation of these candidate biomarkers in larger scale studies. The proposed mechanisms of anti-PD-1 action on Tregs subsets are summarized in **Figure 2**.

## **3.2 Effect of CTLA-4 blockade on regulatory T cells**

Several observations suggest the action of anti-CTLA-4 on Tregs to be key in mediating treatment effects on the tumor. Indeed, some studies involving murine tumor models reported anti-CTLA-4 treatment to simultaneously induce an increased expansion of peripheral Tregs and a decreased expansion of intra-tumoral Tregs [82–84]. This dual action of anti-CTLA-4 on Tregs proliferation could be due to the higher expression of CTLA-4 on exhausted tumor-infiltrating Tregs, where their decline under treatment is suggested to be mediated by a distinct Fc-gamma receptor dependent mechanism of anti-body-mediated cell-mediated cytotoxicity (ADCC)

**35**

*Evolving Dynamic Biomarkers for Prediction of Immune Responses to Checkpoint Inhibitors…*

[85]. This feature was suggested to play a key role in tumor control under CTLA-4 blockade since it has been demonstrated that an on-treatment increase in intratumoral Teffs: Tregs ratio stands as the correlate of an optimal treatment response [82, 83, 86, 87]. Moreover, in another study in different murine models, the therapeutic activity of ipilimumab was found to essentially rely on this Fc-dependent Tregs depletion and not on the checkpoint inhibitor action of the drug [88]. However, there is noticeable inconsistencies in observed Tregs dynamics under CTLA-4 blockade in humans. Indeed, certain studies document a remarkable expansion of Tregs in the peripheral blood of patients treated with anti-CTLA-4 [51, 89–91], while other studies report a declined or unchanging frequency of peripheral Tregs during this therapy [37, 49, 92]. In addition, the ability of the peripheral Tregs dynamics to predict a treatment response to anti-CTLA-4 treatment is also quite unclear. For example, their change in frequency is found to correlate both negatively [93] and positively [90] with anti-CTLA-4 treatment benefit. Yet some other studies show no correlation at all between the change in Tregs frequency and a treatment response to anti-CTLA-4 therapy [56, 94, 95]. Observations of the intra-tumoral Tregs dynamics under anti-CTLA-4 is also inconclusive. A contradictory effect is put forth by a cohort on regionally advanced melanoma patients treated with 2 neoadjuvant doses of ipilimumab. This study reported a reversed association between the change in intra-tumoral Tregs frequency and treatment benefit [90]. Similarly, another study reported a marked decline in intra-tumoral Tregs levels in melanoma patients responding to ipilimumab compared to non-responding ones [96]. On the contrary, two other studies report increasing frequencies of Tregs in biopsies of patients undergoing anti-CTLA-4 therapy [13, 32]. It is interesting to note that the accumulation of Tregs within the tumors upon CTLA-4 blockade may be induced by a feedback loop triggered by a successful cytotoxic T cell response [97]. This may account for the positive correlation between the intra-tumoral levels of Tregs and patient long-term survival as reported by some studies involving solid tumors [98, 99]. These observations nonetheless suggest that Tregs dynamics under CTLA-4 treatment, in the TME and possibly in the periphery, should be accounted for when monitoring for treatment effects. The depletion of intra-tumoral Tregs under CTLA-4 blockade is illustrated in **Figure 2**.

**3.3 Effect of PD-1 and CTLA-4 blockade on myeloid cell compartment**

Monocytes, macrophages and dendritic cells are involved in antigen presentation and T cells priming and thereby serve as a bridge between the innate and adaptive immune response. However, chronic inflammation arising due to cancer disturbs the myeloid cell line maturation process, leading to the generation of myeloid derived suppressor cells [MDSCs] and tumor-associated macrophages [TAMs], that are both suppressors of the anti-tumor immune response [100]. These tumor associated monocytes and macrophages are known to display a wide variety of phenotypes with both pro-inflammatory [M1] and immunosuppressive [M2] functions [101]. Likewise, several studies on animal models have found that treatment with ICIs has the ability of bringing about a spectacular transformation of the intra-tumoral myeloid cell compartment from an immunosuppressive configuration to a more pro-inflammatory one [102, 103]. It has been suggested that the increased INF-γ secretion by renewed T cells would possibly indirectly mediate this myeloid cell reprogramming in the TME under immune checkpoint therapy [102]. Also, it was observed that dual PD-1 and CTLA-4 blockade induces an increase in intra-tumoral pro-inflammatory macrophages, as shown in animal models [104]. In addition, potential direct mechanisms of regulation of MDSCs by anti-PD-1 or anti-CTLA-4 treatment were also identified. As an example, an induced expression of CTLA-4 on monocyte-derived dendritic cells [mDCs] acts as a negative

*DOI: http://dx.doi.org/10.5772/intechopen.96494*

**Figure 2.** *Effect of immune checkpoint inhibition on immune suppressive T cells.*

*Evolving Dynamic Biomarkers for Prediction of Immune Responses to Checkpoint Inhibitors… DOI: http://dx.doi.org/10.5772/intechopen.96494*

[85]. This feature was suggested to play a key role in tumor control under CTLA-4 blockade since it has been demonstrated that an on-treatment increase in intratumoral Teffs: Tregs ratio stands as the correlate of an optimal treatment response [82, 83, 86, 87]. Moreover, in another study in different murine models, the therapeutic activity of ipilimumab was found to essentially rely on this Fc-dependent Tregs depletion and not on the checkpoint inhibitor action of the drug [88]. However, there is noticeable inconsistencies in observed Tregs dynamics under CTLA-4 blockade in humans. Indeed, certain studies document a remarkable expansion of Tregs in the peripheral blood of patients treated with anti-CTLA-4 [51, 89–91], while other studies report a declined or unchanging frequency of peripheral Tregs during this therapy [37, 49, 92]. In addition, the ability of the peripheral Tregs dynamics to predict a treatment response to anti-CTLA-4 treatment is also quite unclear. For example, their change in frequency is found to correlate both negatively [93] and positively [90] with anti-CTLA-4 treatment benefit. Yet some other studies show no correlation at all between the change in Tregs frequency and a treatment response to anti-CTLA-4 therapy [56, 94, 95]. Observations of the intra-tumoral Tregs dynamics under anti-CTLA-4 is also inconclusive. A contradictory effect is put forth by a cohort on regionally advanced melanoma patients treated with 2 neoadjuvant doses of ipilimumab. This study reported a reversed association between the change in intra-tumoral Tregs frequency and treatment benefit [90]. Similarly, another study reported a marked decline in intra-tumoral Tregs levels in melanoma patients responding to ipilimumab compared to non-responding ones [96]. On the contrary, two other studies report increasing frequencies of Tregs in biopsies of patients undergoing anti-CTLA-4 therapy [13, 32]. It is interesting to note that the accumulation of Tregs within the tumors upon CTLA-4 blockade may be induced by a feedback loop triggered by a successful cytotoxic T cell response [97]. This may account for the positive correlation between the intra-tumoral levels of Tregs and patient long-term survival as reported by some studies involving solid tumors [98, 99]. These observations nonetheless suggest that Tregs dynamics under CTLA-4 treatment, in the TME and possibly in the periphery, should be accounted for when monitoring for treatment effects. The depletion of intra-tumoral Tregs under CTLA-4 blockade is illustrated in **Figure 2**.

#### **3.3 Effect of PD-1 and CTLA-4 blockade on myeloid cell compartment**

Monocytes, macrophages and dendritic cells are involved in antigen presentation and T cells priming and thereby serve as a bridge between the innate and adaptive immune response. However, chronic inflammation arising due to cancer disturbs the myeloid cell line maturation process, leading to the generation of myeloid derived suppressor cells [MDSCs] and tumor-associated macrophages [TAMs], that are both suppressors of the anti-tumor immune response [100]. These tumor associated monocytes and macrophages are known to display a wide variety of phenotypes with both pro-inflammatory [M1] and immunosuppressive [M2] functions [101]. Likewise, several studies on animal models have found that treatment with ICIs has the ability of bringing about a spectacular transformation of the intra-tumoral myeloid cell compartment from an immunosuppressive configuration to a more pro-inflammatory one [102, 103]. It has been suggested that the increased INF-γ secretion by renewed T cells would possibly indirectly mediate this myeloid cell reprogramming in the TME under immune checkpoint therapy [102]. Also, it was observed that dual PD-1 and CTLA-4 blockade induces an increase in intra-tumoral pro-inflammatory macrophages, as shown in animal models [104]. In addition, potential direct mechanisms of regulation of MDSCs by anti-PD-1 or anti-CTLA-4 treatment were also identified. As an example, an induced expression of CTLA-4 on monocyte-derived dendritic cells [mDCs] acts as a negative

*Advances in Precision Medicine Oncology*

of specific CD8+

function. Therefore, 4PD1Hi cells are considered as a marker of tumor progression. Anti-PD-1 treatment was shown to reduce the proliferation of 4PD-1Hi cells, whereas anti-CTLA-4 treatment showed an exactly opposite effect on these cells. In line with this interesting observation, the downregulation of 4PD-1Hi cells under anti-PD-1 treatment was further documented as a biomarker of treatment response under anti-PD-1 pembrolizumab antibody in a melanoma patient cohort. In addition, some preclinical studies support the fact that anti-PD-1 therapy can induce the expansion

showed that PD-1 blockade in sub-optimally primed T cell conditions supported the

high levels of CD38 and this effect was associated with treatment failure and tumor resistance in cancer patients [81]. The dynamics of these T cell subsets under treatment may therefore provide valuable predictive information, pending the validation of these candidate biomarkers in larger scale studies. The proposed mechanisms of

Several observations suggest the action of anti-CTLA-4 on Tregs to be key in mediating treatment effects on the tumor. Indeed, some studies involving murine tumor models reported anti-CTLA-4 treatment to simultaneously induce an

increased expansion of peripheral Tregs and a decreased expansion of intra-tumoral Tregs [82–84]. This dual action of anti-CTLA-4 on Tregs proliferation could be due to the higher expression of CTLA-4 on exhausted tumor-infiltrating Tregs, where their decline under treatment is suggested to be mediated by a distinct Fc-gamma receptor dependent mechanism of anti-body-mediated cell-mediated cytotoxicity (ADCC)

proliferation of dysfunctional immunosuppressive CD8+

**3.2 Effect of CTLA-4 blockade on regulatory T cells**

*Effect of immune checkpoint inhibition on immune suppressive T cells.*

anti-PD-1 action on Tregs subsets are summarized in **Figure 2**.

T cells immunosuppressive subsets [30]. Indeed, a recent study

T cells expressing PD-1 and

**34**

**Figure 2.**

regulator of mDCs-associated cytokine secretion and antigen-specific CD4+ T cell proliferation [105]. Moreover, subsets of intra-tumoral MDSCs that express PD-1 and CTLA-4 are found to display decreased arginase 1 expression and activity upon anti-CTLA-4 or anti-PD-1 treatment in mice [106]. In murine models, it has been shown that arginase 1 impairs T cells functions and contributes to immune evasion [107]. Furthermore, it has been recently reported that anti-PD-1 therapy was able to prevent the block in myeloid cell lineage maturation, thereby allowing the myeloid precursors to maturate into effector macrophages and dendritic cells contributing favorably to the anti-tumoral immune response [108].

A decline in circulating MDSCs under anti-CTLA-4 is found to correlate with patient outcome in several studies [37, 90, 109, 110], although this association is not universally reported [111]. Moreover, these studies showed discordant observations regarding the dynamics and predictive value of the major MDSCs subsets (monocytic MDSCs (mo-MDSCs) and polymorphonuclear MDSCs (PMN-MDSCs) subsets) [112]. In addition, several studies showed that anti-PD-1 treatment had no effect on the level of circulating mo-MDSC and PMN-MDSC subsets [9, 113, 114]. Yet a particular study revealed a prominent restructuration of the myeloid compartment after initiation of anti-PD-1 therapy in metastatic melanoma patients when studied

**37**

survival [93, 123].

*Evolving Dynamic Biomarkers for Prediction of Immune Responses to Checkpoint Inhibitors…*

**4. Dynamic predictive and prognostic soluble biomarkers in cancer** 

**4.1 Blood cell counts/ratios, C-reactive protein and lactate dehydrogenase**

In the subsections below, we will focus on blood-based candidate biomarkers that can be utilized as predictive or prognostic markers in cancer immunotherapy.

Changes in the blood cell counts and their ratios including neutrophils, lymphocytes, neutrophil to lymphocyte ratio as well as C-Reactive Protein (CRP) and Lactate Dehydrogenase (LDH) have been reported as prognostic/predictive outcome markers for immunotherapy [2]. Several studies have shown that low neutrophils and high lymphocytes are associated with overall survival (OS) in cancer patients [58, 115, 116]. For example, melanoma patients on nivolumab treatment having absolute lymphocyte counts of >1000/μL and absolute neutrophil count of <4000/μL were observed to have better overall survival [115]. On the other hand, pre-treatment neutrophil-to-lymphocyte ratio (NLR) and derived NLR (dNLR) can also serve as an index of the systemic inflammatory response and therefore considered as useful indicators of response in immunotherapy. Pre-treatment NLR/ dNLR levels and survival association studies in advanced cancers including melanoma, non-small-cell lung cancer (NSCLC) and genitourinary have reported that high pre-treatment NLR and dNLR levels are associated with poor progression free survival (PFS)/OS with increased risks of death in immunotherapy treated patients indicating their usefulness as predictive and prognostic biomarkers [117–120]. CRP is an inflammatory marker that induces the expression of acute-phase proteins such as neutrophils and has been correlated with poor prognosis in several cancers [121, 122]. With regards to immunotherapy, post treatment increased CRP levels have been associated with inflammation, disease progression and in some cases immune-related adverse events. On the other hand, low CRP levels post immunotherapies have been associated with better antitumor response/

LDH is a final enzyme in the glycolysis pathway that catalyzes the interconver-

The fact that blood cell counts/ratios, CRP and LDH tests are performed as part of a routine diagnosis and also are highly assessable/measurable at various treat-

ment timelines in patients making them attractive dynamic biomarkers.

sion of pyruvate and lactate. In cancers, high levels of LDH leads to increased utilization of glycolysis as their energy requirement in the microenvironment [124]. Studies have confirmed that LDH is a significant negative prognostic factor for immunotherapy treated stage 4 melanoma patients [125]. Elevated baseline LDH in melanoma and lung cancer patients treated with pembrolizumab/nivolumab is associated with poor OS and higher risk of death [126–128]. Similar results have been reported for advanced esophageal squamous cell carcinoma patients treated with the anti-PD-1 immune checkpoint inhibitor camrelizumab where elevated

LDH levels were found to correlate with poor OS [129].

effect on myeloid cells compartment is illustrated in **Figure 3**.

under the lens of high dimensional single cell analysis platforms [113]. Therefore, the ability to monitor the evolution of myeloid cells under immune checkpoint blockade appears to be of great predictive value, considering the important role that this cell compartment possibly plays in the modulation of the anti-tumoral immune response by either promoting or preventing the effector T cells response observed upon these therapies. One example of the described mechanisms of anti-PD-1 and anti-CTLA-4

*DOI: http://dx.doi.org/10.5772/intechopen.96494*

**immunotherapy**

**Figure 3.** *Effect of PD-1 and CTLA-4 blockade on myeloid cells.*

*Evolving Dynamic Biomarkers for Prediction of Immune Responses to Checkpoint Inhibitors… DOI: http://dx.doi.org/10.5772/intechopen.96494*

under the lens of high dimensional single cell analysis platforms [113]. Therefore, the ability to monitor the evolution of myeloid cells under immune checkpoint blockade appears to be of great predictive value, considering the important role that this cell compartment possibly plays in the modulation of the anti-tumoral immune response by either promoting or preventing the effector T cells response observed upon these therapies. One example of the described mechanisms of anti-PD-1 and anti-CTLA-4 effect on myeloid cells compartment is illustrated in **Figure 3**.
