Prospective primary human lung cancer tumoroids to predict treatment response

The 5-year survival of advanced lung cancer is between 1-5% and there is an
urgent need for new therapeutic options that improve survival and quality of
life. While precision medicine now enables the identification of driver
mutations in tumors, treatments with *targeted* small molecules or antibodies
lead to responses that are nearly always not durable and come with serious
adverse effects in a sizable minority of the patients. Also in immune therapy,
responses are only observed in a minority of patients, and although they may
last for many years, resistance is still the rule. The main reason for
resistance is that clonal evolution and selection of resistant cancer cells
during treatment cannot be predicted prospectively in patients. Currently no
method exists which enables the prediction of durable responses prior to a
specific treatment and therapy modification during relapse. The prediction
would not only be of significant value for individual patients for they would
not be exposed to ineffective and costly therapy, but it would give insight in
the molecular mechanisms for resistance and hence could lead to new ways or
targets to overcome this.

One of the most important barriers to achieve durable responses in advanced
lung cancer is intra- and inter-tumor heterogeneity, a common feature of human
solid cancers. Tumor heterogeneity is thought to be driven by a subpopulation
of tumor cells termed lung cancer initiating cells or lung cancer stem cells
that reflect the *cell of origin* and maintain self-renewal and multipotent
properties of these cells but that are transformed.
Tumoroid technology has enabled the culturing of normal and transformed *stem
cells* directly from patients without any genetic manipulation (i.e. IPS) (1).
Such normal and cancer tumoroids maintain many of the properties of the tumors
and are thought to be an excellent in vitro 3D model system.
Only recently patient-derived tumoroidshave been exploited to establish
prospective tumor tissue banks that can be used for drug screening (1). In our
lab we have successfully established primary 2D and 3D cell culture systems,
tumoroids, including organoids from the proximal bronchus coming from
lobectomies. We are using these systems to predict normal tissue complication
to combination treatments.

We and others have demonstrated that lung stem cell pathways such as the NOTCH
signaling pathway is frequently deregulated in lung cancers and is associated
with a worse outcome. In vitro and in preclinical models deregulation of the
NOTCH pathway is associated with resistance to radiotherapy and first-line
chemotherapy (2). Thus, blocking the NOTCH pathway may improve treatment
response. Clinical trials using NOTCH inhibitors have unfortunately been
unsuccessful in part because of limited response and lack of biomarkers for
patient selection.
Checkpoint inhibitors have changed the outcome of patients with metastatic
non-small cell lung cancer (NSCLC) in first and later lines, with improved
progression-free survival (PFS), overall survival (OS) and quality of life (3).
Radiotherapy has consistently been shown to activate key elements of the immune
system that are responsible for resistance for immune therapy (4-8). Radiation
upregulates MHC-class I molecules that many cancer cells lack or only poorly
express, tumor-associated antigens, provokes immunogenic cell death, activates
dendritic cells, decreases regulatory T-cells (Tregs) in the tumor, broadens
the T-cell repertoire and increases T-cell trafficking, amongst many other
effects. Radiation may convert a completely or partly poorly or non-immunogenic
tumor immunogenic. Radiotherapy in combination with different forms of immune
therapy such as anti-PD-(L)1, anti-CTLA4, immunocytokines, dendritic cell
vaccination and Toll-like receptor agonists improved consistently local tumor
control and very interestingly, lead to better systemic tumor control (the
*abscopal* effect) and the induction of specific anti-cancer immunity with a
memory effect. Moreover, as PD1/PD-L1 is upregulated by radiation and radiation
can overcome resistance for PD-(L)1 blockage, their combination is logical (9).
The best timing, sequencing and dosing of all modalities is a matter of intense
research, but in pre-clinical models, the concurrent administration of
anti-PD-(L)1 was superior to sequential (8). The recently published subgroup
analysis of the phase 1 KEYNOTE-001 trial at a single institution, aimed to
investigate if prior radiotherapy would affect the PFS or the OS (10,11). In
patients having received prior extra-cranial radiotherapy, the six-month PFS
rate was 54.3% vs. 21.4% among never irradiated patients. The median OS was
11.6 months and the six-month OS estimate was 75.3% among patients who
previously received extra-cranial radiation therapy vs. a median OS of 5.3
months and a six-month OS estimate of 45.3% among patients who did not receive
extra-cranial radiation therapy. Although in pre-clinical models the best way
for combining radiation with anti-PD-(L)1 is to give it concurrently or at
least very close to each other (8), in this study, radiotherapy was delivered
in median 9.5 months prior to the first cycle of pembrolizumab. Still, a
beneficial effect may have occurred.
Supporting an enhancing effect of radiotherapy on the immune system in
combination with pembrolizumab, patients with prior thoracic radiotherapy had
more overall pulmonary toxicity compared to never irradiated patients: 12.5%
vs. 1.4%.
It is clear that radiotherapy may well become an integral part of immune
therapy against cancer. Nevertheless, as with all treatments, optimal
biomarkers for response are lacking. They would not only allow patient
selection, but would also give insight in resistance mechanisms and the
identification of new targets or the optimal use of current medications and
radiation, such as dosing and sequencing. Moreover, not only biomarkers for
tumor response, but also for side effects are needed, for the latter may be
dose-limiting and result in the omission of therapy in the more frail and older
patient population. Putative biomarkers for immune response are those
associated with immunogenic cell death (ICD) (12-15). These and other
immune-related markers are currently evaluated within a clinical trial at
Maastro.

Tumoroids are generated from tissue biopsies, and are a collection of
organ-specific cell types that are able to self-organize in-vitro in a manner
similar to the in-vivo situation (3D). They have the capability to facilitate
in-depth analysis of patient*s own tumor material at point of diagnosis and
during progressive/recurrent disease. There is currently no published protocol
to establish short and long-term lung cancer tumoroids from lung cancer
patients. Such a methodology would enable the prospective identification of
*patient tailored optimal treatments* as well as the derivation of predictive
biomarkers for response and relapse.

For this study, we will also be using data, derived from the non WMO trial:
*Prospective primary human lung cancer organoids to predict treatment
response*. This trial was approved by the Ethics Committee of Zuyderland
Hospital in Heerlen (trial number 17-N-139). Surgically obtained tumor tissue
samples have predominantly been used for the optimization of the current
protocol for the establishment of long tumor organoid cultures. Data of this
trial that will be used for further validation and optimization of the current
protocol include the definitive diagnosis and FFPE material to compare the
morphology of the cultured 3D tumor structures with the morphology of the
primary tumor.

To establish matched normal and primary human lung cancer organoids from
patient-derived lung (tumor) material.

Preclinical study, using patient derived lung (tumor) material to establish
tumoroids.

From patients who will ondergo a bronchoscopy during standard care, left-over
tumor tissue will be derived during this standard procedure as well as 10ml of
extra blood through an existing bloodline.