Longitudinal profiling identifies co-occurring BRCA1/2 reversions, TP53BP1, RIF1 and PAXIP1 mutations in PARP inhibitor-resistant advanced breast cancer

BACKGROUND

Resistance to therapies that target homologous recombination deficiency (HRD) in breast cancer limits their overall effectiveness. Multiple, preclinically validated, mechanisms of resistance have been proposed, but their existence and relative frequency in clinical disease are unclear, as is how to target resistance.

PATIENTS AND METHODS

Longitudinal mutation and methylation profiling of circulating tumour (ct)DNA was carried out in 47 patients with metastatic BRCA1-, BRCA2- or PALB2-mutant breast cancer treated with HRD-targeted therapy who developed progressive disease-18 patients had primary resistance and 29 exhibited response followed by resistance. ctDNA isolated at multiple time points in the patient treatment course (before, on-treatment and at progression) was sequenced using a novel >750-gene intron/exon targeted sequencing panel. Where available, matched tumour biopsies were whole exome and RNA sequenced and also used to assess nuclear RAD51.

RESULTS

BRCA1/2 reversion mutations were present in 60% of patients and were the most prevalent form of resistance. In 10 cases, reversions were detected in ctDNA before clinical progression. Two new reversion-based mechanisms were identified: (i) intragenic BRCA1/2 deletions with intronic breakpoints; and (ii) intragenic BRCA1/2 secondary mutations that formed novel splice acceptor sites, the latter being confirmed by in vitro minigene reporter assays. When seen before commencing subsequent treatment, reversions were associated with significantly shorter time to progression. Tumours with reversions retained HRD mutational signatures but had functional homologous recombination based on RAD51 status. Although less frequent than reversions, nonreversion mechanisms [loss-of-function (LoF) mutations in TP53BP1, RIF1 or PAXIP1] were evident in patients with acquired resistance and occasionally coexisted with reversions, challenging the notion that singular resistance mechanisms emerge in each patient.

CONCLUSIONS

These observations map the prevalence of candidate drivers of resistance across time in a clinical setting, information with implications for clinical management and trial design in HRD breast cancers.

Haemodynamic response to induction of anaesthesia with ketamine/midazolam

The haemodynamic responses following induction of anaesthesia with ketamine and midazolam have not been determined previously. Twenty adult patients for elective myocardial revascularization were randomized to two regimens for induction of anaesthesia. Patients in Group I received ketamine, 2 mg.kg-1, and midazolam, 0.2 mg.kg-1 and those in Group II received ketamine, 2 mg.kg-1, and midazolam, 0.4 mg.kg-1. Measurements were recorded at baseline, 1 min post-induction, and at one, three, five and ten minutes after tracheal intubation. Tachycardia and hypertension (greater than 20% increases from awake baseline values) were treated with esmolol, 250 micrograms.kg-1. There were 11 patients in Group I and nine patients in Group II. There were no significant intergroup differences in demographic or haemodynamic variables. Both groups had decreases (P less than 0.05), in stroke volume, pulmonary capillary wedge pressure, and right ventricular end-diastolic volume at multiple study intervals following anaesthetic induction. None of these changes required clinical intervention. Five patients (all in Group II) had hypertensive responses to tracheal intubation. Preoperative hypertension (mean arterial pressure greater than or equal to 100 mmHg) was a predictor (P less than 0.05) of a hypertensive response to intubation, independent of the midazolam dose. Intravenous ketamine and midazolam was associated with a high incidence (25%) of haemodynamic responses to tracheal intubation. The higher dose of midazolam did not provide any haemodynamic advantage.

Anti-kidney antibody teratogenesis: evidence for a direct action on the rat conceptus by a mechanism involving the yolk sac

IgG from anti-rat kidney antiserum (anti-K) was added to 48-h cultures of 9.5-day whole rat conceptuses. At concentrations of 25 micrograms/ml and above, dysmorphogenesis of embryos (optic vesicle defects) was observed. Significant embryonic growth retardation was apparent at 75 micrograms/ml and above. Other abnormalities (neural tube defects, deficits in axial rotation, reduced somite numbers) were seen at 100 and 150 micrograms/ml. In separate experiments, anti-K inhibited pinocytosis of 125I-polyvinylpyrrolidone by isolated yolk sacs. All of these effects of anti-K are similar to those produced by anti-visceral yolk sac antiserum suggesting a similar mechanism of teratogenic action, namely interference with yolk-sac-mediated embryotrophic nutrition.

Failure of zinc to prevent dysmorphogenesis of cultured rat conceptuses by anti-yolk sac antiserum

Day 10 rat conceptuses were cultured for 48h in the presence of either cadmium or anti-visceral yolk sac antiserum (AVYS). Cadmium was embryotoxic at concentrations exceeding 0.25 micrograms/ml whilst AVYS caused embryonic dysmorphogenesis, particularly affecting the optic vesicles, at concentrations of 2 microliters/ml and above. The effect of pretreatment with zinc on embryotoxicity caused by cadmium or AVYS was studied. Zinc ameliorated the effects of cadmium but had no effect on AVYS-induced embryonic abnormalities. In a second set of experiments inhibition of 125I-labelled PVP uptake by the yolk sac of cultured whole conceptuses was studied. Cadmium and AVYS both inhibited uptake compared to control cultures. Zinc again ameliorated the effect of cadmium but had no action against AVYS-induced inhibition. These results are in contrast to our previous findings using isolated cultured yolk sacs in which zinc ameliorated the inhibitory effects on 125I-labelled PVP uptake of both cadmium and AVYS. These data show that in experiments using the isolated cultured yolk sac and the intact cultured conceptus, a qualitatively different response in yolk sac behaviour is observed under similar experimental conditions.

Differential effect of zinc on teratogen-induced inhibition of pinocytosis by cultured rat yolk sac

Pinocytosis as measured by the uptake of 125I labelled PVP by the isolated cultured day 12 rat yolk sac was observed to be linear over a 4 h incubation period and to proceed at a rate of approximately 2.5 microliters/mg protein/h. Cadmium, anti-visceral yolk sac antibody (AVYS) and trypan blue all inhibited pinocytosis in a concentration-dependent fashion when added to the culture medium, although at low concentrations trypan blue was slightly stimulatory. The effect of zinc on the inhibition of pinocytosis by these three teratogens was studied. It was observed that zinc ameliorated the inhibitory effects of cadmium and AVYS, but had no effect on inhibition by trypan blue. These results indicate that the previously demonstrated protective action of zinc against cadmium-induced yolk sac dysfunction is not specific to that agent but extends to inhibition of pinocytosis by AVYS, and further suggest that, because of its refractoriness to zinc, trypan blue-induced inhibition of pinocytosis by yolk sac occurs by a mechanism different from that effected by cadmium and AVYS.

In-vitro teratogenicity of retinoids

Mid-gestation rat conceptuses were cultured for 48 h in serum containing the retinoids all-trans-retinoic acid (TRA), 13-cis-retinoic acid (13-CRA), etretinate (ETR), etretin or one of six retinamides at concentrations ranging from 0.5 to 400 micrograms/ml. TRA was toxic at a concentration of 0.5 microgram/ml. 13-CRA and etretin caused abnormal development at 1.0 microgram/ml. However, the six retinamides were less toxic and adverse developmental effects were only evident at concentrations of 50 or 100 micrograms/ml. ETR was without effect at 100 micrograms/ml, the highest dose level of this compound tested. In vivo, TRA, 13-CRA and ETR are highly teratogenic. In this culture system, TRA and 13-CRA caused abnormal development at very low concentrations but in contrast, ETR was non-toxic at 100 micrograms/ml. Therefore these findings indicate that in vivo, maternal pharmacokinetics, and bioactivation in particular, play a major role in inducing abnormal development. Cis/trans isomerization was not a major determinant of toxicity. However, there appeared to be a relationship between abnormal development and the actual or estimated pKa values of the 10 retinoids tested.