The Intrauterine Device: How to Deploy this Strategy in the Molecular World?

Summary:

Progestin-based therapy can safely be offered to a subset of patients with atypical endometrial hyperplasia or Grade 1 endometrioid endometrial cancer who desire fertility preservation. A recent study shows that levonorgestrel intrauterine device confers durable clinical benefit and identifies possible immune mechanisms of relapse and resistance.

In this issue of Clinical Cancer Research, Bowen and colleagues (1) report long-term follow-up data from their prospective Phase 2 clinical trial of levonorgestrel intrauterine device (LIUD) as a non-operative treatment option for women with atypical endometrial hyperplasia (AEH) or Grade 1 endometrioid endometrial cancer (EC). They provide important updates on two fronts. First, they determined that nearly 40% of patients who initially responded to LIUD eventually suffered disease relapse, confirming that, for many individuals, continuous progestin therapy should not be viewed as a definitive alternative to hysterectomy. Second, they leveraged digital spatial transcriptomic profiling to demonstrate that the tumor immune microenvironment within relapsed lesions is characterized by features of immune exhaustion. This lends insight into potential mechanisms of resistance to progesterone-based therapy and, in turn, helps advance our understanding of which individuals might be best poised for durable response.

Endometrial cancer is the most common gynecologic malignancy in the United States. Most cases are diagnosed at an early stage when it is still confined to the uterus and potentially curable. The current standard of care for individuals with AEH or suspected locoregional EC is complete surgical resection. There are few non-surgical options for premenopausal women who desire fertility preservation or for those in whom surgery is high-risk due to medical comorbidities, but treatment with continuous progestin-based therapy can be offered to a subset of well-selected patients with AEH or stage 1 endometrioid EC. These lesions are thought to arise in the setting of chronic excess estrogen exposure without the counterbalance of sufficient opposing progesterone. Synthetic progestins appear to exert their anti-tumor effect by activating progesterone receptors and downregulating estrogen-driven signaling. Multiple studies have shown that treatment with continuous progestin therapy can yield high response rates, but the long-term management strategies remain less well defined, as do the factors that predict progesterone response and resistance.

The data presented by Bowen et al. (1) come from the first longitudinal follow-up of a prospective trial of LIUD and therefore constitute an important milestone for the field, especially with regards to the relatively high rate of relapse. What then are the oncofertility implications for patients who receive continuous progestin therapy and respond to treatment? The high rate of and variable time to relapse presents inherent complications for reproductive planning. As testament to that complexity, only one of the ten patients in this study who attempted to conceive was ultimately able to achieve a spontaneous pregnancy. A second patient elected for embryo cryopreservation, and two others unfortunately suffered progression of disease while undergoing in vitro fertilization. These outcomes underscore the ongoing need for consensus, evidence-based guidelines to help direct preconception, antepartum, and peripartum care for this specialized group of patients.

Identifying biomarkers of progesterone response and resistance will help further optimize patient selection for treatment and monitoring. Bowen and colleagues (1) make important strides to this end. The authors performed a thorough investigation of potential clinicodemographic predictors of long-term response. Interestingly, there was no significant relationship between the duration of response and body mass index (BMI), which perhaps comes as a surprise given the known relationship between obesity, immune exhaustion, and cancer risk, and given the authors’ observation that relapsed lesions exhibit features of immune exhaustion. Younger age was associated with shorter duration of response in univariate analysis, and while it was not significant in a separate multivariable analysis, it remains worth noting given the population for whom fertility-sparing treatment is most desirable; 65.6% of premenopausal participants from the study had reported their desire for future childbearing.

Bowen and colleagues (1) take an important next step by seeking to define the molecular markers of response and resistance to LIUD. Their multi-faceted approach included an investigation into the baseline expression of numerous estrogen-regulated genes. This revealed that pre-treatment elevation of insulin like growth factor (IGF)-1 and IGF-2 were both significantly associated with a longer duration of treatment response (Figure 1), an observation that is particularly interesting considering prior studies showing that IGF signaling is associated with endometrial hyperplasia and tumorigenesis, and that progesterone might down-regulate IGF1 signaling in uterine leiomyoma. (2, 3) Taken together, these data suggest that endometrial carcinomas dependent on IGF signaling may be particularly sensitive to progesterone, and the predictive utility of IGF1/2 expression should therefore be prospectively validated.

Figure 1.

Patients with atypical endometrial hyperplasia or grade 1 endometrioid endometrial cancer received treatment with the levonorgestrel intrauterine device as a potential fertility-sparing therapeutic approach. Elevated baseline expression of IGF-1 and IGF-2 was associated with a longer duration of response. On-treatment lesions exhibited an increase in NK cells and cytotoxic lymphocytes. Relapsed lesions were characterized by increased IDO1 expression, reduced NK cell abundance, and increased IFNα and TGFβ signaling.

Perhaps their most exciting observations come from longitudinal digital spatial transcriptomic profiling of the tumor immune microenvironment. This was performed in a subset of 5 patients who had both pre- and on-treatment tissue available for analysis; in 3 of the patients, post-relapse tissue was also analyzed. This assessment revealed that on-treatment tissue is enriched with infiltrating NK cells and cytotoxic lymphocytes compared to pre-treatment tissue, which suggests that both innate and adaptive immunity play a role in the initial anti-tumor immune response. Disease relapse appears to then occur in the setting of immune escape and was marked by a significant increase in IDO1 expression. IDO1 exhibits immunosuppressive properties and has been shown to inhibit NK and cytotoxic lymphocyte proliferation (Figure 1). (4)

In line with this, Bowen and colleagues (1) found that post-relapse tissue contained significantly fewer NK cells as well as decreased cytotoxic lymphocytes compared to on-treatment tissue. The shift in the tumor immune microenvironment at the time of relapse is further characterized by increased type 1 interferon (IFNα) and TGFβ signaling, which suggests an immune exhausted landscape. Outside of the field of oncology, progesterone has been broadly implicated in facilitating immune dysregulation by downregulating inflammatory reactions and the activation of immune cells. Further questions to explore include: 1) which specific cells overexpress IDO1 in response to continuous progestin; 2) what are the expression dynamics of other immune checkpoints such as PD-1, PD-L1, CTLA-4, and LAG-3; and 3) whether there is a causal relationship between progesterone and immune exhaustion?

Lastly, we now recognize that Grade 1 EC is molecularly heterogenous; while the majority are copy number-low per TCGA classification, up to 40% can be mismatch repair deficient (dMMR) or POLE mutated. (5) We know from prior studies that MMRd EC tends to respond poorly to continuous progestin therapy. (6) Thus, the possibility that progesterone may modulate the tumor environment and impact subsequent response to PD-1 blockade should be carefully considered. Additional data in regard to TCGA classification and LIUD outcomes will be essential to successfully select patients for whom LIUD is unlikely be successful so they can proceed promptly to curative surgery.

Overall, the data presented by Bowen and colleagues (1) show that continuous progestin therapy remains an effective treatment option for many patients with AEH and early-stage EC. Further characterization of post-treatment relapse patterns, biomarkers of response and resistance, and the impact of progesterone on the tumor microenvironment and immune milieu are needed to better guide long-term surveillance strategies and oncofertility care.