From bedside to bench and back

Getting a diagnosis

For many patients, bladder cancer begins with a shock in the bathroom: blood in the urine. It may be painless. It may appear only once and then disappear. But it can mark the beginning of a long journey through examinations, tissue sampling, treatment and uncertainty. It normally leads to an investigation that includes cystoscopy – a camera examination of the inside of the bladder – and, when a tumor is found, surgery to remove tissue for diagnosis. The pathology result establishes how far the tumor has grown and whether it has features associated with a greater risk of returning or spreading. The answer can change the course of the patient’s life.

For some patients, the tumor has not yet grown into the muscular wall of the bladder. The prognosis may be relatively favorable, but treatment is not always a single event followed by a return to normal life. Given the high recurrence rate of bladder cancer, patients often require ongoing surveillance and repeated interventions, including cystoscopic examinations, urinary monitoring, surgical management, and intravesical therapies. Many people with bladder cancer remain under surveillance for a decade or for the rest of their lives. Every appointment can bring back the same question: has the cancer returned?

Patients and their families may need to absorb complex information while also confronting uncertainty about whether a proposed treatment will work.

For others, the disease is already more serious when it is discovered. The cancer may have invaded the bladder muscle or spread to other parts of the body. Decisions about major surgery and drug treatment can then become urgent. Patients and their families may need to absorb complex information while also confronting uncertainty about whether a proposed treatment will work.

A major disease burden

Bladder cancer is the ninth most common cancer worldwide. More than 600,000 people were diagnosed in 2022, and more than 220,000 died from the disease. It is also considered one of the most challenging and costly cancers to diagnose and manage, partly because many patients require repeated invasive examinations.

Approximately 75% of tumors have not yet invaded the bladder muscle when they are diagnosed, while 25% are already muscle-invasive.

In Sweden, approximately 3,200 people are diagnosed, and around 700 die from bladder cancer each year. More than 30,000 people were living with a diagnosis of urothelial cancer at the end of 2022. The median age at diagnosis is 76, and around three times as many men as women develop the disease. Approximately 75% of tumors have not yet invaded the bladder muscle when they are diagnosed, while 25% are already muscle-invasive.

Treatment options

For many years, chemotherapy – drugs that attack rapidly dividing cells – was the principal drug treatment for advanced bladder cancer. It can kill cancer cells and shrink tumors, but it also affects healthy cells and can cause substantial side effects. Doctors now have access to treatments that work in more targeted ways. One approach is immunotherapy, which helps the patient’s immune system recognize and attack the cancer. Tumors can use molecular signals to apply a kind of brake to immune cells; immunotherapy is intended to release that brake.

Another approach uses an antibody as a delivery vehicle (antibody-drug conjugates, ADCs). The antibody recognizes a particular protein on the surface of the cancer cell and carries a powerful cell-killing drug toward it. The aim is to concentrate more of the treatment within the tumor, although healthy tissues can still be affected. Combinations of these two approaches have produced significant and transformative improvements for patients with previously untreated advanced bladder cancer.

An unanswered clinical question

Bladder cancer is a heterogeneous disease that can exhibit highly variable clinical courses, largely reflecting the underlying biology of each affected individual, explains Anders Ullén, professor of clinical oncology at Karolinska Institutet and senior physician at Karolinska University Hospital.

Some patients achieve lasting remission, while others experience limited benefit, despite having tumors that appear histologically similar. “The goal is to tailor and individualize treatment according to the specific characteristics and circumstances of each patient,” he says. 

In recent years, several novel and highly effective systemic oncological therapeutic options have become available; however, predictive molecular biomarkers are still lacking for many of these treatments. 

“The development and implementation of new diagnostic approaches that can improve our ability to predict treatment efficacy prior to therapy initiation would be of substantial value, both for patients and for healthcare systems,” says Ullén.

Clinicians face a difficult question: how can they identify the treatment most likely to help a particular patient before valuable time is lost, while also minimizing toxicity?

At present, they have limited ability to do so. A patient may undergo months of demanding treatment before scans reveal that it is not working. During that period, the disease may progress, side effects may reduce the patient’s quality of life, and the opportunity to try another treatment may become narrower. That uncertainty was the starting point for Charlotte Stadler and Anders Ullén’s ongoing research collaboration.

Mapping the tumor’s cellular neighborhood

Charlotte Stadler, docent in spatial proteomics at KTH Royal Institute of Technology, heads SciLifeLab’s Spatial Proteomics Unit and conducts translational cancer research in close collaboration with clinicians at Karolinska Institutet and Region Stockholm. Together with Anders Ullén, she is investigating whether the organization and functional state of cells inside a tumor can help predict treatment response. 

Conventional pathology often examines one protein marker at a time. A tissue section may, for example, be assessed for a protein called PD-L1. The protein is part of a natural system that prevents the immune response from becoming overactive. Cancer cells can exploit this system to protect themselves from attack. PD-L1 can bind to another protein, PD-1, on the surface of immune cells. When the two connect, they send an inhibitory signal – a molecular instruction telling the immune cell to slow down. Immunotherapies can block this interaction and allow the immune response to become active again. Simply finding PD-L1 in a tumor, however, does not necessarily show that the molecular brake is being applied. Nor does it reliably reveal whether a patient will respond to treatment.

Spatial proteomics provides a more detailed view. It can identify several proteins in the same tissue section while retaining information about which cells express them, where those cells are located, and how they relate to one another.

Stadler and Ullén’s teams have developed a workflow that combines multiplexed immunofluorescence (mIF) with an in situ proximity ligation assay, or isPLA, developed by the Swedish company Navinci. This technique uses antibodies and fluorescent signals to identify several types of cells and proteins in one tissue section. The Navinci assay adds another layer: it detects whether PD-1 and PD-L1 are sufficiently close to enable physical interaction.

Stadler and Ullén’s teams have developed a workflow that combines multiplexed immunofluorescence (mIF) with an in situ proximity ligation assay, or isPLA, developed by the Swedish company Navinci.

In an initial proof-of-concept analysis, the researchers studied tissue from patients with metastatic bladder cancer who had received immunotherapy. They combined PD-1/PD-L1 proximity analysis with a 14-marker antibody panel, allowing them to identify the immune and tumor cells associated with each proximity signal.

This enabled the researchers to connect the molecular readout to specific immune-cell populations. The clearest association involved cytotoxic CD8-positive T cells – immune cells capable of killing cancer cells. PD-1/PD-L1 proximity signals associated with these cells were enriched among patients who experienced a complete response, whereas they were uncommon among patients whose disease progressed.

“Many patients are treated with immunotherapy, and particularly one that targets the PD-1/PD-L1 signaling between cells,” says Stadler.

“While this has been a game-changer for some patients, many do not benefit from the therapy despite expression of PD-L1. This can lead to side effects and toxicity, while paying for expensive therapies with no benefit,” she says. “We wanted to find a better predictive marker to know who really would benefit from this therapy. It makes sense to look at whether the signaling is present in the tumor, so that there is something to block by giving the therapy. Using the combination of isPLA with mIF, we can be more confident that the relevant signaling is there, and what cell types are involved.”

Translation begins with a clinical problem

The bladder cancer work is one tangible example of SciLifeLab’s Translation to Healthcare initiative. Translation is sometimes presented as the final stage of research: a discovery is made in an academic laboratory and is later transferred to a hospital or company. In this project, the clinical problem comes first. Karolinska University Hospital contributes patient material, pathology expertise, treatment information, and an understanding of the decisions clinicians face. SciLifeLab contributes assay development, advanced imaging infrastructure, and computational analysis.

This bidirectional exchange is the principle behind both “bench to bedside” and “bedside back to bench.” Technology development is guided by the conditions of healthcare from the beginning, rather than being optimized in isolation and confronted with clinical reality only at the end.

The tissue is part of the technology

A tumor sample may consist of a small biopsy, material from an operation that cannot be repeated, or an archival block that has already supported the patient’s diagnosis and several previous tests. Sample handling is one of the major barriers to clinical translation, particularly when very small biopsies must support several forms of molecular analysis. Before a protein can be measured or a clinician can act on the result, the material must be collected, processed, and preserved in a way that protects its biological information.

Multiplexed imaging often involves repeated cycles of antibody staining, signal generation, and antibody removal. Some established methods use high-temperature treatment between cycles. The heat removes the antibodies used in one round so that a new set can be applied. The process can work very well, but tissue tolerance varies. Repeated heating can alter sensitive protein targets, affect tissue morphology, or cause parts of the section to detach from the microscope slide.

Stadler’s group is therefore evaluating AtlasPlex, developed by Stockholm-based Atlas Antibodies, as a possible complement to the Navinci assay. The idea is to use isPLA first to detect the PD-1/PD-L1 interaction and then add a small panel of markers to identify the tumor and immune cells surrounding it, all in the same tissue section. The potential advantage lies in how the staining cycles are performed. Instead of using repeated antibody-removal cycles, AtlasPlex uses a heat-free chemical step that inhibits the enzyme responsible for producing the fluorescent signal. This allows the next staining round to begin without repeatedly heating the tissue.

“The integration of functional assays such as isPLA with mIF can be done with various detection methods and not all are suitable for all tissue samples. A general challenge with integration of different assays or in multiplexing is to find conditions that allow for robust signal detection without compromising tissue quality. Retained tissue morphology is also crucial for the next generation of pathological analysis,” explains Stadler.

A general challenge with integration of different assays or in multiplexing is to find conditions that allow for robust signal detection without compromising tissue quality.

“We are therefore interested in finding protocols that are as gentle as possible on the samples while giving a satisfying signal for the targeted markers. Protocols that allow for detecting several markers without using high temperatures have the potential to also work on more fragile samples and to better preserve structures and detailed morphology.”

For the patient, the journey may begin with blood in the urine and a small piece of tissue removed during an examination. The goal of translation is to carry the information in that sample all the way back to the patient – ultimately helping clinicians make treatment decisions that are more precise, more timely, and more likely to help.