Quick answer
A team at the Chinese PLA General Hospital analysed 61 tumour samples from 34 people with advanced kidney cancer and describes a chain of signals that appears to drive resistance to the standard first-line combination of a tyrosine kinase inhibitor (a targeted drug) and an immune checkpoint inhibitor. In people whose tumours did not respond, tumour cells producing a protein called SAA recruited a subtype of neutrophil — a common immune cell — which in turn carried a protein, CEACAM1, that switches off the CD8+ T cells that would otherwise attack the tumour [1]. In mouse models, blocking SAA or interrupting the CEACAM1–TIM-3 interaction restored the effect of the drugs [1][2].
Key takeaways
- 61 tumour samples from 34 patients with advanced renal cell carcinoma were profiled, yielding 333,169 single cells after quality control [1].
- People who responded had more CD8+ T cells in their tumours; people who did not respond had more neutrophils, and neutrophil levels tracked with poorer outcomes [1].
- A VEGFA+CEACAM1+ neutrophil subset was identified as the key immunosuppressive population; it expands after treatment specifically in non-responders [1].
- The proposed chain runs SAA+ tumour cells → VEGFA+ neutrophils → CEACAM1 → TIM-3 on CD8+ T cells → T-cell exhaustion [1][2].
- In mice, a triple combination adding an SAA-blocking antibody to the standard two drugs worked better than the standard pair; separately, blocking CEACAM1–TIM-3 restored anti-PD-1 activity [1].
- The authors call 34 patients modest and say prospective validation is needed before any of this reaches the clinic [1].
The condition in context: advanced kidney cancer
Kidney cancer, and specifically renal cell carcinoma, accounts for roughly 2–3% of adult cancers worldwide. About 30% of patients already have metastatic disease when they are diagnosed, and 20–40% of people whose tumour appears localised later relapse after surgery. Once the disease is advanced, the five-year survival rate is 11.7% [1].
| Region | What the data shows | Source |
|---|---|---|
| Worldwide | 434,840 new kidney cancer cases and 155,953 deaths in 2022; the 14th most common cancer. | World Cancer Research Fund, from GLOBOCAN 2022 [3] |
| China | 73,656 new cases in 2022 — the largest number of any country — and 23,991 deaths. Age-standardised rate 3.1 per 100,000. | WCRF / GLOBOCAN 2022 [3] |
| United States | 71,759 new cases in 2022 at an age-standardised rate of 12.8 per 100,000. The American Cancer Society's estimates for 2026 are about 80,450 new cases and 15,160 deaths. | WCRF / GLOBOCAN 2022 [3]; ACS Cancer Facts & Figures 2026 [4] |
| Japan | 21,207 new cases in 2022, age-standardised rate 6.9 per 100,000. | WCRF / GLOBOCAN 2022 [3] |
China has the largest number of cases but a low age-standardised rate (3.1 per 100,000) compared with the United States (12.8). That gap is mostly about population size and age structure rather than about risk, and it is why raw case counts and age-standardised rates should not be read as the same kind of number [3].
What the study did
The team, led within the urology department of the Chinese PLA General Hospital — Zhang Xu, Wang Baojun, Ma Xin, Ci Weimin and Huang Yan among the corresponding authors, with Gu Liangyou as first author — collected tumour tissue from patients before and after treatment with a tyrosine kinase inhibitor alone or combined with an immune checkpoint inhibitor [1][5].
They then read the genetic activity of individual cells. In plain terms: rather than grinding up a tumour and averaging it, single-cell sequencing lets you see which cell types are present and what each is doing. Combining their own samples with two published datasets, they profiled 333,169 cells and compared people whose tumours shrank with people whose tumours did not [1].
Two differences stood out. Responders' tumours were rich in CD8+ T cells — the immune cells that kill tumour cells. Non-responders' tumours were rich in neutrophils, and neutrophils scored lowest of any immune cell type for predicting a good response. Within the neutrophils, one subtype carrying VEGFA and CEACAM1 expanded after treatment only in the non-responders [1].
They then traced the chain backwards: tumour cells making SAA (an acute-phase inflammatory protein) appeared more often in non-responders and increased further after treatment; SAA was shown to raise CEACAM1 on neutrophils in the laboratory; and CEACAM1 binds TIM-3, a receptor on exhausted CD8+ T cells. Blocking either end of that chain improved drug activity in mouse models [1].
The team also measured SAA in blood plasma taken before treatment, and found higher levels in the patients who went on not to respond [1]. A blood test that flags likely resistance before months of expensive treatment would answer the exact problem clinicians describe: patients who get no benefit but still carry the side effects and the cost. The authors are careful that this needs validation in larger prospective cohorts [1].
Why it matters, and what it does not settle
| What it tells you | What it does not tell you |
|---|---|
| There is now a named, testable mechanism for resistance to the standard first-line combination in advanced kidney cancer [1]. | Whether drugs aimed at SAA or CEACAM1 work in people. The evidence so far is in mice. |
| Neutrophils, not just T cells, appear central to resistance — a shift from the usual T-cell-centred explanation [1][2]. | How much of resistance this single axis explains. Other routes almost certainly exist. |
| A blood-based signal (plasma SAA) separated responders from non-responders in this cohort [1]. | Whether that test is accurate enough to guide treatment. 34 patients is too few to know. |
Availability and limits
- This is laboratory research, not a treatment. Nothing here is available to patients. The SAA and CEACAM1–TIM-3 interventions were tested in mice.
- 34 patients. The authors state plainly that the cohort is appropriate for single-cell work but modest for generalising, and that larger prospective cohorts are needed [1].
- Paired samples were hard to get. Biopsy feasibility and treatment-induced necrosis limited how many patients had both before and after tissue [1].
- Mouse results often do not carry over to people. The authors note possible interspecies differences [1].
- Standard care has not changed. TKI plus immunotherapy remains first line; this work does not alter what a patient should be offered today.
What to ask your oncologist
| Question | Why it matters |
|---|---|
| Which first-line combination are you recommending, and why that one for me? | Regimens differ in side-effect profile and evidence base. |
| How will we know early whether it is working? | Imaging intervals and what counts as progression should be agreed in advance. |
| Are there trials of resistance mechanisms or biomarker-led selection that I could enter? | This is how findings like this one get tested in people. |
| What is the plan if the first combination does not work? | Second-line options exist and are worth knowing about early. |
Questions patients ask
What does "resistance" mean here?
Two things can happen: the treatment never works (primary resistance) or it works and then stops (acquired resistance). This study looked at both, and reports that 5–20% of patients fail to respond to the standard combination at all [1].
What are neutrophils doing in a tumour?
Neutrophils are the immune system's first responders to infection and injury. In tumours they can be reprogrammed to suppress other immune cells. Here, a specific VEGFA+CEACAM1+ subtype appeared to switch off the CD8+ T cells that attack the cancer [1].
Is there a blood test for this now?
Not one you can order. The team measured SAA in pre-treatment plasma and found higher levels in eventual non-responders, but the authors say this needs validation in larger prospective cohorts before it could be used clinically [1].
Can I get a treatment that targets this pathway?
No. The SAA-blocking antibody and the CEACAM1–TIM-3 blocking agent in this paper were used in mouse models. Neither is an approved treatment for kidney cancer [1].
Does this mean my current treatment is the wrong one?
No. TKI plus immunotherapy remains the standard first-line treatment and helps most people. This work is about the minority who do not benefit, and it is a hypothesis for future trials rather than a change to today's practice [1][4].